ASGE guideline on screening and surveillance of Barrett's ...

27
GUIDELINE ASGE guideline on screening and surveillance of Barretts esophagus Prepared by: ASGE STANDARDS OF PRACTICE COMMITTEE Bashar Qumseya, MD, MPH, FASGE, 1, * Shahnaz Sultan, 2, * Paul Bain, 3 Laith Jamil, MD, FASGE, 4 Brian Jacobson, 5 Sharmila Anandasabapathy, 6 Deepak Agrawal, MD, MPH, MBA, 7 James L. Buxbaum, MD, FASGE, 8 Douglas S. Fishman, MD, FASGE, 9 Suryakanth R. Gurudu, MD, FASGE, 10 Terry L. Jue, MD, FASGE, 11 Sapna Kripalani, MD, 12 Jeffrey K. Lee, MD, 13 Mouen A. Khashab, MD, 14 Mariam Naveed, MD, 15 Nirav C. Thosani, MD, 16 Julie Yang, MD, 17 John DeWitt, 18 Sachin Wani, MD, FASGE, ASGE Standards of Practice Committee Chair 19 The final document was approved by the ASGE Governing Board and the Standards of Practice Committee and represents the official guideline of the ASGE on these topics. This document is the ofcial American Society for Gastrointestinal Endoscopy guideline on screening and surveillance in patients with Barretts esophagus (BE) and is based on systematic reviews of the evidence using the Grading of Recommendations, Assessment, Develop- ment and Evaluation methodology. The document ad- dresses key clinical questions that include the role and impact of screening and surveillance and the utility of advanced imaging and sampling modalities like chro- moendoscopy, volumetric laser endomicroscopy, confocal laser endomicroscopy, EUS, and wide-area transepithelial sampling. This guideline complies with the standards for guideline development set forth by the Institute of Medicine for the creation of trustworthy guide- lines and aims to help clinicians understand the pub- lished literature and the quality of available data with the ultimate goal of optimizing care for patients. This American Society for Gastrointestinal Endoscopy (ASGE) guideline addresses screening and surveillance of Barretts esophagus (BE) using the Grading of Recommen- dations, Assessment, Development and Evaluation (GRADE) methodology. 1 BE, a premalignant condition for esophageal adenocarcinoma (EAC), is characterized by the replacement of the normal squamous epithelium of the distal esophagus with metaplastic intestinal-type columnar epithelium. 2 BE is diagnosed in 7% to 10% of individuals with chronic GERD and is estimated to be present in 1% to 2% of the general adult population. 3,4 The incidence of EAC continues to be among the fastest- rising incidence cancers in the Western population and has been closely mirrored by a rise in EAC-related mortal- ity. 5,6 In the United States, an estimated 16,940 persons were diagnosed with esophageal cancer (>60% with EAC) in 2017, and 15,690 died from their disease. 5,7,8 Prog- nosis for patients with EAC is strongly related to the stage at diagnosis. Unfortunately, most patients with EAC are diagnosed with late-stage disease, and the overall 5-year survival is <20%, which decreases to <5% for patients with distant disease at diagnosis. 3,5 Although the effective- ness of treatment strategies in the management of EAC has improved over the last decade, 9,10 recent data from the U.S. National Cancer Institutes Surveillance, Epidemiology, and End Results program suggest that the proportion of EAC cases with localized, regional, or distant stage disease remains relatively stable, and, strik- ingly, 40% of EAC patients are still diagnosed with distant disease. 5 BE progresses to invasive EAC through stages of intesti- nal metaplasia (nondysplastic BE [NDBE]), low-grade dysplasia (LGD), to high-grade dysplasia (HGD), to intra- mucosal carcinoma, and nally to invasive EAC. 11 It is this step-wise progression in patients with BE and stage- dependent survival in EAC that provides the impetus for screening and surveillance of BE. To ultimately impact the morbidity and mortality associated with EAC, several national and international medical societies recommend screening for BE in individuals with multiple risk factors and surveillance when the diagnosis of BE is estab- lished. 4,12-14 The last decade has seen several advances that attempt to address the limitations of current screening and surveillance strategies. These include construction of models to identify individuals at risk for BE and EAC and those at the highest risk of progression when diagnosed with BE, noninvasive and less expensive tools for screening, and the use of advanced imaging and sampling techniques during endoscopy. In this guideline document, we address the best available evidence for screening and surveillance of BE. Our previous guideline, published in 2018, addressed issues related to endoscopic eradication therapy (EET) for patients with BE-associated dysplasia and intramucosal cancer. 2 www.giejournal.org Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 335

Transcript of ASGE guideline on screening and surveillance of Barrett's ...

Page 1: ASGE guideline on screening and surveillance of Barrett's ...

w

GUIDELINE

ww.giejournal.org

ASGE guideline on screening and surveillance of Barrett’sesophagus

Prepared by: ASGE STANDARDS OF PRACTICE COMMITTEE

Bashar Qumseya, MD, MPH, FASGE,1,* Shahnaz Sultan,2,* Paul Bain,3 Laith Jamil, MD, FASGE,4

Brian Jacobson,5 Sharmila Anandasabapathy,6 Deepak Agrawal, MD, MPH, MBA,7

James L. Buxbaum, MD, FASGE,8 Douglas S. Fishman, MD, FASGE,9 Suryakanth R. Gurudu, MD, FASGE,10

Terry L. Jue, MD, FASGE,11 Sapna Kripalani, MD,12 Jeffrey K. Lee, MD,13 Mouen A. Khashab, MD,14

Mariam Naveed, MD,15 Nirav C. Thosani, MD,16 Julie Yang, MD,17 John DeWitt,18

Sachin Wani, MD, FASGE, ASGE Standards of Practice Committee Chair19

The final document was approved by the ASGE Governing Board and the Standards of Practice Committee andrepresents the official guideline of the ASGE on these topics.

This document is the official American Society for at diagnosis. Unfortunately, most patients with EAC are

Gastrointestinal Endoscopy guideline on screening andsurveillance in patients with Barrett’s esophagus (BE)and is based on systematic reviews of the evidence usingthe Grading of Recommendations, Assessment, Develop-ment and Evaluation methodology. The document ad-dresses key clinical questions that include the role andimpact of screening and surveillance and the utility ofadvanced imaging and sampling modalities like chro-moendoscopy, volumetric laser endomicroscopy,confocal laser endomicroscopy, EUS, and wide-areatransepithelial sampling. This guideline complies withthe standards for guideline development set forth by theInstitute of Medicine for the creation of trustworthy guide-lines and aims to help clinicians understand the pub-lished literature and the quality of available data withthe ultimate goal of optimizing care for patients.

This American Society for Gastrointestinal Endoscopy(ASGE) guideline addresses screening and surveillance ofBarrett’s esophagus (BE) using the Grading of Recommen-dations, Assessment, Development and Evaluation(GRADE) methodology.1 BE, a premalignant conditionfor esophageal adenocarcinoma (EAC), is characterizedby the replacement of the normal squamous epitheliumof the distal esophagus with metaplastic intestinal-typecolumnar epithelium.2 BE is diagnosed in 7% to 10% ofindividuals with chronic GERD and is estimated to bepresent in 1% to 2% of the general adult population.3,4

The incidence of EAC continues to be among the fastest-rising incidence cancers in the Western population andhas been closely mirrored by a rise in EAC-related mortal-ity.5,6 In the United States, an estimated 16,940 personswere diagnosed with esophageal cancer (>60% withEAC) in 2017, and 15,690 died from their disease.5,7,8 Prog-nosis for patients with EAC is strongly related to the stage

diagnosed with late-stage disease, and the overall 5-yearsurvival is <20%, which decreases to <5% for patientswith distant disease at diagnosis.3,5 Although the effective-ness of treatment strategies in the management ofEAC has improved over the last decade,9,10 recent datafrom the U.S. National Cancer Institute’s Surveillance,Epidemiology, and End Results program suggest thatthe proportion of EAC cases with localized, regional, ordistant stage disease remains relatively stable, and, strik-ingly, 40% of EAC patients are still diagnosed with distantdisease.5

BE progresses to invasive EAC through stages of intesti-nal metaplasia (nondysplastic BE [NDBE]), low-gradedysplasia (LGD), to high-grade dysplasia (HGD), to intra-mucosal carcinoma, and finally to invasive EAC.11 It isthis step-wise progression in patients with BE and stage-dependent survival in EAC that provides the impetus forscreening and surveillance of BE. To ultimately impactthe morbidity and mortality associated with EAC, severalnational and international medical societies recommendscreening for BE in individuals with multiple risk factorsand surveillance when the diagnosis of BE is estab-lished.4,12-14 The last decade has seen several advancesthat attempt to address the limitations of current screeningand surveillance strategies. These include construction ofmodels to identify individuals at risk for BE and EAC andthose at the highest risk of progression when diagnosedwith BE, noninvasive and less expensive tools forscreening, and the use of advanced imaging and samplingtechniques during endoscopy. In this guideline document,we address the best available evidence for screening andsurveillance of BE. Our previous guideline, published in2018, addressed issues related to endoscopic eradicationtherapy (EET) for patients with BE-associated dysplasiaand intramucosal cancer.2

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 335

Page 2: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

GENERAL CONSIDERATIONS FOR SCREENINGAND SURVEILLANCE

Screening is the mechanism through which populationsmay be assessed to identify individuals who have a diseaseor a preclinical condition that predisposes to a disease. Sur-veillance is the program through which these at-risk indi-viduals are periodically assessed or examined to identifydisease at a stage amenable to cure.15 The evidence tosupport screening and surveillance endoscopy, a practicethat has been recommended by national guidelines,4,12-14

is highly variable and based on observational data. As dis-cussed later, no randomized controlled trials (RCTs) haveevaluated these 2 practices.

The ideal study to assess the effectiveness of screeningand surveillance is an RCT of individuals at risk for BE andEAC who are randomized to undergo screening upperendoscopy (EGD) compared with no screening. Surveil-lance would then be performed on individuals diagnosedwith BE, and the frequency of surveillance would be basedon grade of dysplasia. Patients with HGD/intramucosal car-cinoma and select cases with LGD would undergo EET(ablation and/or endoscopic resection techniques) andthose with invasive cancer esophagectomy.2 Patients withNDBE would undergo surveillance endoscopy every 3 to5 years. The primary outcome of this study would becomparison of EAC mortality (critical endpoint), andsecondary outcomes would include all-cause mortality,EAC incidence, stage of diagnosis, number of EAC patientsundergoing esophagectomy, and stage of diagnosis. Astudy designed in this fashion would provide the bestevidence for the clinical practice of screening and surveil-lance to achieve the overall aim of reducing mortalityrelated to EAC. However, such a study would require thou-sands of subjects with decades of follow-up, given the lowincidence of cancer in patients with BE.15 Such studies donot exist at this time, and to our knowledge, no suchstudies have been funded or are underway. It isimportant to recognize the risk of bias in publishedstudies that falsely increase the apparent benefit ofsurveillance practices. Lead-time bias is the detection ofasymptomatic preclinical cancer by endoscopy, whichmay only increase the detection time of individuals withcancer without truly increasing life-years. Length-timebias occurs when slowly growing cancers are more likelydetected during endoscopy than rapidly growing cancers,which artificially creates the appearance that screeningand surveillance prolongs survival. These considerationsare important to understand in our effort to make mean-ingful clinical recommendations.

AIMS AND SCOPE

The aim of this document is to offer evidence-based rec-ommendations and clinical guidelines addressing key issues

336 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

related to screening and surveillance in patients with BE.The panel considered the following clinical questions:� What is the role of surveillance endoscopy in patients

with NDBE compared with no surveillance in decreasingthe rate of cancer progression, EAC-related mortality,and all-cause mortality?

� What is the role of screening for BE in the general pop-ulation and at-risk populations compared with noscreening in decreasing the rate of cancer progressionand overall mortality?

� What is the role of advanced imaging technologies inimproving the detection of dysplasia in BE patients un-dergoing surveillance?o The advanced imaging techniques addressed in thisdocument include chromoendoscopy (CE), includingvirtual chromoendoscopy (VC), confocal laser endo-microscopy (CLE), and volumetric laser endomicro-scopy (VLE).

� What is the role of wide-area transepithelial sampling(WATS) in improving dysplasia detection rates?

� What is the role of EUS in staging BE patients withdysplasia and early EAC?

METHODS

OverviewThis guideline document was based on systematic re-

views (SRs) of the available literature for each clinical ques-tion. The quality or certainty in the evidence and strengthof recommendations was based on the GRADE framework.When existing SRs were identified, they were used toinform the guideline when appropriate. If no existing SRswere found, a new SR (and meta-analysis [MA], whenpossible) was conducted with the help of an expertlibrarian. Evidence profiles were created with the help ofGRADE methodologists (S.S., B.Q.), and recommendationswere drafted by the panel at a Standards of Practicemeeting convened in Nashville on March 17, 2018. ForPopulation, Intervention, Comparator, and Outcomes(PICO) 2 and PICO 6a, a conference call including thefull GRADE panel was conducted to readdress the clinicalquestion and recommendations based on updated pub-lished data. Throughout the document, white-light endos-copy (WLE) with targeted biopsy sampling from all visibleabnormalities and random 4-quadrant biopsy samplingevery 1 to 2 cm starting from the top of the gastric foldsup to the most proximal extent of the BE, otherwisereferred to as the Seattle Protocol, was used as the crite-rion standard for each of the relevant clinical questions ad-dressing surveillance and use of advanced imagingtechniques to increase the yield of dysplasia detection.4,16

Panel composition and conflict of interestmanagement

The panel composition consisted of content experts(J.D., S.A., B.Q., S.W.), GRADE methodologists (S.S., B.Q.),

www.giejournal.org

Page 3: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

patient representative, primary care physician (S.K.), healthpolicy expert (B.J.), and members of the Standards of Prac-tice committee. All members were asked to disclose con-flicts of interests based on the ASGE policy (https://www.asge.org/forms/conflict-of-interest-disclosure and https://www.asge.org/docs/default-source/about-asge/mission-and-governance/asge-conflict-of-interest-and-disclosure-policy.pdf). Panel members who received funding for any technol-ogies or companies associated with any of the PICOs wereasked to declare this before the discussion and did not voteon the final recommendation addressing that specific PICOquestion.

Formulation of clinical questionsThese topics were conceptualized by the authors of the

documents and members of the ASGE Standards of Practicecommittee and approved by the Governing Board. Thequestions followed the PICO format: P, population in ques-tion; I, intervention; C, comparator; and O, outcomes of in-terest. For all clinical questions, potentially relevant patient-relevant outcomes were identified a priori and rated fromnot important to critical through a consensus process. Rele-vant clinical outcomes included cancer-specific and all-causemortality, progression to EAC, dysplasia detection rates, andperformance characteristics of diagnostic tests. A list of thePICO questions is detailed in Table 1.

Literature search and study selection criteriaFor each PICO question, either existing SRs/MAs were

identified and reviewed or a new SR/MA was conducted.Details of the search strategies are reported in Appendix1 (available online at www.giejournal.org) as perPreferred Reporting Items for Systematic Reviews andMeta-analyses criteria. An expert medical librarian per-formed all searches. Citations were imported into EndNote(Thompson Reuters, Philadelphia, Pa), and duplicates wereremoved. The EndNote library was uploaded into Covi-dence (www.covidence.org), and 2 independent reviewerswere assigned to each search. Each study was reviewedbased on title and abstract using explicit inclusion andexclusion criteria. If applicable, full text was then reviewed.Differences with resolved by consensus.

Data extraction and statistical analysisWhen necessary, data extraction was accomplished by 2

reviewers using Microsoft Excel (Microsoft Corporation,Redmond, Wash). The outcomes varied by PICO andincluded increased diagnostic yield (relative and absolute),pooled sensitivity and specificity, pooled relative risk (RR),odds ratio (OR), or proportions (risk of BE). For outcomeswith limited or no available direct comparisons, indirectcomparisons were used to estimate the magnitude and di-rection of effect. We assessed heterogeneity using the I2

and Q statistic. We used random effects models for most an-alyses, and studies were weighted based on their size. Pub-lication bias was assessed using funnel plots and the classic

www.giejournal.org

fail-safe. Statistical analyses were performed using Compre-hensive Meta Analysis V3 (Biostat Inc, Englewood, NJ).

For diagnostic performance, we used 2 prevalence esti-mates to illustrate the number of true positives, true neg-atives, false positives, and false negatives. Theseestimates were derived from content experts using thebest current available evidence. The estimate of 5% repre-sents the prevalence of BE patients with or withoutdysplasia in patients referred for EGD in the community,and the estimate of 30% represents the prevalence of BEpatients with or without dysplasia in patients referred forEGD at tertiary care or referral centers.

Certainty in evidence (quality of evidence)The certainty of the evidence was determined using the

GRADE framework, which starts with defining the health-care question in terms of the population of interest, alterna-tive management strategies (intervention and comparator),and all patient-important outcomes.17 A systematic search isthen performed to identify all relevant studies, and datafrom individual included studies are used to generate anestimate of the effect for each patient-important outcomeas well as a measure of the uncertainty associated withthat estimate (typically a confidence interval [CI]). TheGRADE approach to rating the quality or certainty of evi-dence begins with the study design (RCTs or observationalstudies) and then addresses 5 reasons to possibly rate downthe quality of evidence (methodologic limitations, inconsis-tency, indirectness, imprecision, and publication bias) and3 reasons to possibly rate up the quality (large effect, dose–response gradient, plausible confounding). The final qualityof evidence ranges from very low to high (Table 2).Guideline developers then formulate the recommendation(s)and consider the direction (for or against) and grade thestrength (strong or weak) of the recommendation(s) basedon the criteria outlined in the GRADE approach. The GRADEevidence profile, developed using GDTpro application(http://gdt.guidelinedevelopment.org/app), contains detailedinformation about the quality of evidence assessment and thesummary of findings for each of the included outcomes.

Considerations in the development ofrecommendations

The strength of a recommendation reflects the extent towhich a guideline panel is confident that desirable effects ofan intervention outweigh undesirable effects, or vice versa,across the range of patients for whom the recommendationis intended. GRADE specifies 2 categories of the strength ofa recommendation: strong or conditional. The main factorsthat drive the recommendation include balance betweenbenefits and harms, taking into account the best estimatesof the magnitude of effects and importance of outcomes;overall quality of evidence; confidence in values and prefer-ences and their variability; and cost/resource implications.The final wording of the recommendations (including direc-tion and strength), remarks, and qualifications were decided

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 337

Page 4: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 1. List of Population, Intervention, Comparator, and Outcomes questions

Population Intervention Comparator Outcomes Rating

1. NondysplasticBarrett’s esophagus

Surveillance endoscopy(varying intervals)

No surveillance 1. Mortality/survival2. Stage of diagnosis of

esophageal adenocarcinoma3. Time to progression to dysplasia

or esophageal adenocarcinoma

CriticalCritical

Critical

2. Population at riskfor Barrett’s esophagus

Screening for Barrett’sesophagus

No screening 1. Mortality/survival2. Stage of diagnosis of

esophageal adenocarcinoma3. Prevalence of Barrett’s esophagus

based on risk factors

CriticalCritical

Important

3. Patients withBarrett’s esophagusundergoing surveillance

Using chromoendoscopy:1) Chromoendoscopy2) Virtual chromoendoscopy

White-lightendoscopy

with random biopsysampling

1. Increase yield in detection ofdysplasia/neoplasia

2. Performance characteristics(accuracy, sensitivity, specificity, etc.)

Critical

Critical

4. Patients withBarrett’s esophagusundergoing surveillance

Using confocal laserendomicroscopy (endoscope-

based and probe-based)

White-lightendoscopy

with random biopsysampling

1. Increase yield in detection ofdysplasia/neoplasia

2. Performance characteristics(accuracy, sensitivity, specificity, etc.)

Critical

Critical

5. Patients withBarrett’s esophagusand suspected dysplasia

EUS No EUS 1. Increased yield in detection ofdysplasia/neoplasia

2. Performance characteristics(accuracy, sensitivity, specificity, etc.)

Analysis conducted at the T1a vs. T1blevel and at the T1 vs. T2 level

Critical

Critical

6. Patients withBarrett’s esophagusundergoing surveillance

New techniques:1) Volumetric laser endosmicroscopy2) Wide-area transepithelial sampling

White-lightendoscopywith random

biopsy sampling

1. Increased yield in detection ofdysplasia/neoplasia

2. Performance characteristics(accuracy, sensitivity, specificity, etc.)

Critical

Critical

EUS, Endoscopic ultrasound.

ASGE guideline on screening and surveillance of BE

by consensus and were approved by all members of thepanel. According to the GRADE approach, the recommen-dations are either “strong” or “conditional” (Table 3). Thewords “the guideline panel recommends” are used forstrong recommendations and “suggests” for conditionalrecommendations.

RESULTS

The recommendations for each clinical question aresummarized in Table 4.

Question 1: What is the role of surveillanceendoscopy in patients with NDBE compared withno surveillance in decreasing the rate of cancerprogression, EAC-related mortality, and all-causemortality?

Recommendation: In patients with NDBE,we suggest performing surveillance endoscopycompared with no surveillance (conditional recom-mendation, very low quality of evidence).

Summary of the evidence: Surveillance endoscopyfor BE is predicated on the assumption that it detects

338 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

dysplasia or early EAC, which can be treated by minimallyinvasive treatment modalities such as EET and thus re-duces morbidity and mortality related to EAC. Despitethe fact that surveillance endoscopy to evaluate fordysplasia is routinely performed for BE, there are noRCTs demonstrating that surveillance (vs no surveillance)leads to a reduction in the proportion of patients withadvanced-stage presentation of EAC or a reduction inEAC-related or all-cause mortality.

A recent SR by Codipilly et al18 provided acomprehensive overview of the available evidence on theimpact of BE surveillance on survival in BE patientsdiagnosed with EAC. The panel relied on this publishedSR and MA to inform the recommendations for this PICOquestion. In this review, the authors focused on 2 groupsof studies: (1) cohort studies of patients with BEwho had undergone surveillance and BE patients withno surveillance (or inadequate surveillance) and (2)prospective or retrospective cohort studies that comparedEAC patients with a history of BE (diagnosed �6 monthsbefore the diagnosis of EAC) with patients whose initialsymptomatic presentation was EAC, without previouslydocumented BE. The authors performed an MA of the

www.giejournal.org

Page 5: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 2. GRADE categories of quality of evidence

GRADE quality of evidence Meaning Interpretation

High We are confident that the true effect lies closeto that of the estimate of the effect.

Further research is very unlikely to change ourconfidence in the estimate of the effect.

Moderate We are moderately confident in the estimateof the effect; the true effect is likely to be closeto the estimate of the effect, but there is apossibility that it is substantially different.

Further research is likely to have an impact on ourconfidence in the estimate of the effect and may

change the estimate.

Low Our confidence in the estimate of the effect islimited; the true effect may be substantiallydifferent from the estimate of the effect.

Further research is very likely to have an impact on ourconfidence in the estimate of the effect and is likely to

change the estimate.

Very low We have very little confidence in the estimateof the effect; the true effect is likely to be

substantially different from the estimate of theeffect.

Any estimate of the effect is very uncertain.

GRADE, Grading of Recommendations, Assessment, Development and Evaluation.

ASGE guideline on screening and surveillance of BE

cohort studies for the following outcomes: mortality (EAC-related and all-cause), EAC stage at time of diagnosis, andrates of operative intervention. These pooled estimateswere incorporated into the evidence profile (Table 5).

For EAC-related mortality, the authors identified 4 pro-spective cohort studies comparing regular BE surveillance(with incomplete or no surveillance) that provided unad-justed data on EAC-related mortality. MA of these studiesshowed that BE surveillance was associated with a reduc-tion in EAC-related mortality (RR, .60; 95% CI, .50-.71).The authors also identified 4 cohort studies showing thatEAC patients with a prior BE diagnosis had a significantlylower risk of mortality because of EAC (unadjusted RR,.73; 95% CI, .57-.94). Only 1 study adjusted for lead-timebias (using a sojourn time of 3 years) as well as stageand treatment of cancer. In this study by El-Serag et al,19

the association between EAC mortality risk andendoscopic surveillance was no longer found (hazardratio [HR], 1.72; 95% CI, .78-2.07). Another study byTramontano et al20 provided lead-time–adjusted data forEAC-related mortality (sojourn time of 3 years) and adjust-ment for lead-time bias eliminated the association betweensurveillance and mortality in those with a prior BE diag-nosis versus those with no prior BE diagnosis (HR, .89;95% CI, .78-1.01).

For all-cause mortality, MA of 3 studies showed thatreceiving BE surveillance (vs incomplete or no surveillance)was associated with a 25% reduction in mortality (unad-justed HR, .75; 95% CI, .59-.94). Additionally, MA of 12studies showed that in EAC patients, having a prior diagnosisof BE was associated with a reduction in mortality (RR, .48;95% CI, .37-.63). Because this reduction in overall mortalitycould have been because of lead- and length-time bias,the authors performed analyses using lead-time–adjustedestimates (using a sojourn time of 2-2.5 years). The pooledeffect estimate from the 3 studies that provided HRs forsurvival after adjusting for lead-time bias showed an

www.giejournal.org

attenuation in the mortality benefit (HR, .85; 95% CI, .75-.95). This study was unable to provide length-time–adjustedanalyses.

For earlier-stage EAC (stages 0 and 1) at diagnosis, theauthors pooled unadjusted data from 4 studies showingthat patients who had BE surveillance were more likelyto be diagnosed with early-stage EAC as compared withthose in the inadequate or no surveillance group (RR,2.11; 95% CI, 1.08-4.11). Additionally, pooled data from 9studies in EAC patients showed that individuals with priorBE diagnosis were more likely to present with an earlierstage EAC compared with those without a prior BE diag-nosis (RR, 5.52; 95% CI, 3.70-8.24).

For the outcome of adverse events (AE), we used datafrom a previously published guideline on AEs of upperGI endoscopy by the ASGE.21 Although the rate of AEsvaries by procedure, indication, and level of sedation,AEs related to surveillance endoscopy are infrequent. Therate of overall AE in diagnostic EGDs is very low (from 1/200 to 1/10,000). The rate of AEs is higher if EET isperformed.22

Certainty in the evidence: The overall certainty inthe evidence for EAC-related mortality was very low orlow because these studies were observational studieswith heterogeneity (for which we rated down). For all-cause mortality the certainty in the evidence was verylow; we rated down for imprecision and inconsistency.Finally, for earlier stage at diagnosis, we had very low-quality evidence and rated down for inconsistency(Table 5).

Considerations: The panel considered additional evi-dence on cost-effectiveness analyses, patient values andpreferences, and harm. A recent SR focused on the eco-nomic impact of screening and surveillance to reducemortality from EAC was used to guide this discussion.15

Most economic studies on cost-effectiveness of screeningincorporate surveillance among individuals diagnosed

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 339

Page 6: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 3. Interpretation of definitions of strength of recommendation using GRADE framework

Implications for Strong recommendation Conditional recommendation

Patients Most individuals in this situation would want the recommendedcourse of action, and only a small proportion would not.

Most individuals in this situation would want thesuggested course of action, but many would not.

Clinicians Most individuals should receive the intervention.Formal decision aids are not likely to be needed

to help individual patients make decisions consistentwith their values and preferences.

Recognize that different choices will be appropriate forindividual patients and that you must help each patient arriveat a management decision consistent with his or her valuesand preferences. Decision aids may be useful in helping

individuals to make decisions consistent with their values andpreferences.

Policymakers The recommendation can be adopted as policyin most situations. Compliance with this

recommendation according to the guideline couldbe used as a quality criterion or performance indicator.

Policymaking will require substantial debate and involvementof various stakeholders.

GRADE, Grading of Recommendations, Assessment, Development and Evaluation.

TABLE 4. Summary of recommendations and quality of evidence

StatementStrength of

recommendation Quality of evidence

1. In patients with nondysplastic BE, we suggest performing surveillanceendoscopy compared with no surveillance.

Conditional Very low

2. There is insufficient evidence on the effectiveness of screening for BE. However, if screeningendoscopy for BE is performed, we suggest a screening strategy that identifies an at-riskpopulation. An at-risk population is defined as individuals with a family history of EAC or BE(high risk) or patients with GERD plus at least 1 other risk factor (moderate risk).

NA NA

3. In patients with BE undergoing surveillance, we recommend using chromoendoscopy, includingvirtual chromoendoscopy and Seattle protocol biopsy sampling, compared with white-lightendoscopy with Seattle protocol biopsy sampling.

Strong Moderate

4. In patients with BE undergoing surveillance, we suggest against routine use of confocal laserendomicroscopy compared with white-light endoscopy with Seattle protocol biopsy sampling.

Conditional Low

5. In BE patients with high-grade dysplasia/IMC or nodules, we recommend againstroutine use of EUS to differentiate mucosal vs submucosal disease.

Strong Moderate

6a. In patients with known or suspected BE, we suggest using WATS-3D in additionto Seattle protocol biopsy sampling compared with white-light endoscopywith Seattle protocol biopsy sampling.

Conditional Low

6b. In patients with BE undergoing surveillance, there is insufficient evidenceto recommend for or against routine of VLE.

No recommendation NA

BE, Barrett’s esophagus; NA, not applicable; EAC, esophageal adenocarcinoma; IMC, intramucosal cancer; VLE, volumetric laser endomicroscopy; WATS-3D, wide-areatransepithelial sampling with computer-assisted 3-dimensional analysis.

ASGE guideline on screening and surveillance of BE

through screening to account for the beneficial effectsof treatment of BE-related dysplasia and early EAC. Ninestudies have been published that have assessed the cost-effectiveness of screening in BE (6 of these studieswere conducted before the advent of EET), and allwere Markov simulations.23-28 With a willingness-to-paythreshold of <$100,000 (ie, society would be willing topay $100,000 for each quality-adjusted life-year gained),all studies found endoscopic screening and surveillanceof subjects with GERD to be cost-effective (5 studiesshowed that the cost was $4000-$15,000 per quality-adjusted life-year gained). There are limited data on patientvalues and preferences with regard to surveillance

340 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

endoscopy. The panel relied heavily on the views ex-pressed by the patient representative who expressedstrong support for surveillance, placing a high value onearly detection of EAC and potential benefits of treatmentand a low value on any burden or harms associated withrepeat surveillance endoscopies.

Discussion: The panel recognized the limitations inthe current available evidence, and despite the apparentshortcomings, a decision in support of surveillance endos-copy was made (conditional recommendation). The panelacknowledged the many limitations in the existing studies:suboptimal study design, varying surveillance protocolswithin and among studies, and issues around lead- and

www.giejournal.org

Page 7: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

length-time bias that can skew results in favor of surveil-lance. An important driver for this recommendation wasthe potential survival benefit demonstrated in favor of sur-veillance. Results of the Barrett’s Oesophagus SurveillanceStudy trial,29 which is currently underway in the UnitedKingdom, are awaited. This multicenter RCT randomizedBE patients to either an intensive surveillance arm(every 2 years) or a nonsurveillance arm (“at need”surveillance).29

Surveillance intervals are currently determined by thepresence and grade of dysplasia. This PICO questiondoes not address the frequency of surveillance intervalsin patients with NDBE or the ideal approach to surveillancebiopsy sampling. Several studies have demonstrated a lowrate of progression to EAC in NDBE (.1%-.3% per year).30-32

On the basis of these data, previous clinical practice guide-lines recommend surveillance endoscopy every 3 to 5 yearsin this patient population, a recommendation that thispanel endorsed until more evidence is available.4,12,33

The distribution of dysplasia and early EAC is highly focaland variable.34 Because a systematic biopsy samplingprotocol detects more dysplasia and early EAC comparedwith random biopsy sampling,16,35,36 medical societies,including this panel, recommend a biopsy sampling proto-col (Seattle biopsy sampling protocol) that uses 4-quadrantbiopsy sampling at 2-cm intervals in patients withoutdysplasia and at 1-cm intervals in patients with priordysplasia, along with targeted biopsy sampling from anymucosal abnormality.3,4

As currently practiced, endoscopic surveillance of BEhas numerous limitations. Compliance rates with the aboverecommendations in terms of surveillance intervals and ob-taining biopsy specimens using the Seattle protocol arepoor.1,34,37,38 Current surveillance programs are time-consuming, and given the highly focal and variable distri-bution of dysplasia and early EAC, even the most thoroughbiopsy sampling surveillance program has the potential forsampling errors. In addition, there is significant interob-server and intraobserver variability among communityand expert pathologists in the interpretation ofdysplasia.39-41 There is a growing interest in the develop-ment of risk-prediction models to identify patients withNDBE at risk for the development of HGD/EAC. A recentSR and MA identified the following factors associatedwith progression: increasing age (OR, 1.03; 95% CI, 1.01-1.05), male sex (OR, 2.16; 95% CI, 1.84-2.53), ever smoking(current or past; OR, 1.47; 95% CI, 1.09-1.98), increasingBE length (OR, 1.25; 95% CI, 1.16-1.36), and LGD (OR,4.25; 95% CI, 2.58-7.0).42 On the other hand, the use ofproton pump inhibitors (OR, .55; 95% CI, .32-.96) andstatins (OR, .48; 95% CI, .31-.73) have been linked to adecreased risk of progression to dysplasia/neoplasia.Another recent randomized trial43 assessed the benefit ofproton pump inhibitors and aspirin in patients with BE.They reported that high-dose proton pump inhibitorwith aspirin was associated with a significant benefit on a

www.giejournal.org

composite endpoint of all-cause mortality, HGD, and EAC(time ratio, 1.59; 95% CI, 1.14-2.23; P Z .0068).

The use of prediction models using demographic andclinical factors has the potential for improving the effective-ness of current surveillance strategies.44-47 In a recent study,a scoring system based on male sex, smoking, BE length,and baseline LGD was developed that identified patientswith BE at low-, intermediate-, and high-risk groups forHGD/EAC.47 External validation before clinical applicationis required for these described prediction models.

Question 2: What is the role of screening for BEcompared with no screening in decreasing therate of cancer progression, EAC-related mortality,and all-cause mortality? What is the role of ascreening strategy that identifies individuals atrisk of having BE?

Recommendation: There is insufficient evidenceon the effectiveness of screening for BE. However,if screening endoscopy for BE is performed, we sug-gest a screening strategy that identifies an at-riskpopulation, defining “at-risk” individuals as thosewith a family history of EAC or BE (high risk) or pa-tients with GERD plus at least 1 other risk factor(moderate risk).

Summary of the evidence: Screening for BE has beenendorsed by medical societies based on the assumptionthat screening will detect individuals with BE and these in-dividuals will be enrolled in surveillance programs todetect dysplasia and early EAC, who will then undergominimally invasive procedures (EET or esophagectomy)and ultimately reduce the incidence, morbidity, and mor-tality associated with EAC. Screening also has the potentialto identify individuals with prevalent dysplasia and earlyEAC who can be treated with EET. We conducted a system-atic search for studies addressing screening for BE in thegeneral population. We found no studies comparingscreening with no screening in individuals at risk for BE.

We then searched for SRs and MAs assessing various riskfactors for BE to identify at-risk populations who maybenefit from screening. The panel decided to use a cut-off prevalence of 10% for BE in any given group (with orwithout risk factors) to recommend screening. Thefollowing risk factors were assessed: history of GERD,male gender, white race, smoking, obesity, and family his-tory of BE and EAC. For the risk factor of GERD, Tayloret al48 analyzed 26 studies and reported an OR of 2.9(95% CI, 1.86-4.54) for the association of GERD with BE,albeit associated with high heterogeneity (I2 Z 89%).This OR increased to 4.5 in 12 studies with GERD for atleast 2 weeks. For the risk factor of smoking, Andriciet al49 conducted an SR and MA of 10 studies andshowed that the odds of having BE significantly increasein smokers when controlling for possible confounders.For the risk of obesity, Singh et al50 conducted an SRand MA assessing the effect of obesity and central

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 341

Page 8: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 5. Evidence profile for surveillance compared with no surveillance/incomplete surveillance in patients with known Barrett’s esophagus

Certainty assessment

No. of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations

EAC-related mortality (unadjusted)*

4 Observational studies Seriousy Not serious Not serious Not serious None

4 Observational studies Not serious Seriousz Not serious Not serious None

EAC-related mortality (accounting for lead time bias and adjustment for patient factors including stage, and treatment)x1 Observational studies Not serious Not serious Not serious Serious{ None

All-cause mortality (unadjusted)*

12 Observational studies Not serious Not serious Not serious Serious{ None

3 Observational studies Not serious Seriousz Not serious Serious{ None

All-cause mortality adjusted (accounting for lead-time)*

3 Observational studies Not serious Seriousz Not serious Serious{ None

Earlier stage of diagnosis (unadjusted)*

4 Observational studies Not serious Seriousz Not serious Not serious None

9 Observational studies Serious Seriousz Not serious Not serious None

AEs from endoscopy

EAC, Esophageal adenocarcinoma; AE, adverse events; RR, relative risk; HR, hazard ratio.*Based on the SR and MA by Codipilly et al18 these estimates were derived from (1) prospective cohort studies of BE patients who had undergone surveillance compared withBE patients with no or incomplete surveillance and (2) prospective or retrospective cohort studies that compared EAC patients with a history of BE (diagnosed >6 months beforean EAC diagnosis) with patients whose initial symptomatic presentation was EAC without previously documented BE.yThere were concerns about bias based on the Newcastle-Ottowa Scale for most studies.zThere was a large amount of heterogeneity between studies for which we rated down.xBased on the SR and MA by Codipilly et al18 this estimate was derived from the study by El-Serag et al,19 which provided an EAC-mortality estimate that was adjusted for lead-time bias, stage, and treatment of EAC.{We rated down for imprecision.

ASGE guideline on screening and surveillance of BE

adiposity. Based on 11 studies, they found no associationwith obesity but reported a significant association withcentral adiposity (Table 6).

In a more recent SR and MA by Qumseya et al,51 theauthors estimated the prevalence of BE in patients withspecific risk factors who underwent screening and triedto synthesize the interactions between such risk factors.The authors identified 48 studies with over 300,000patients, of which >1900 had biopsy specimen–confirmed BE. The prevalence of BE was assessed in thegeneral or low-risk populations and in those individualswith risk factors such as GERD, family history of BE orEAC, age >50 years, obesity, and male gender. When con-trolling for population Western versus non-Western, meanage, and gender distribution, there was a linear relation-ship between the number of risk factors and the risk ofBE. As the number of risk factors increased, the risk ofBE increased by 1.2% per additional risk factor. In thisstudy, all patients with multiple risk factors had eitherGERD or age >50.

342 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

Based on the above evidence, the panel outlined a high-risk group, a moderate-risk group, and a low-risk group.The panel recognized patients with a family history ofEAC or BE as a high-risk group and recommendedscreening for BE in that population. The panel describeda moderate-risk group to include patients with GERDand at least 1 additional risk factor who may also benefitfrom screening for BE. The additional risk factors wereage >50, obesity/central adiposity, history of smoking, ormale gender. Finally, the panel did not recommendscreening in low-risk patients.

Considerations: The panel again recognized the lackof data on the impact of screening for BE on EAC inci-dence, EAC mortality, or all-cause mortality but acknowl-edged that existing guidelines from other GI societiescurrently endorse screening for BE in specific popula-tions.4 Eight studies were identified that estimated thecost-effectiveness of screening to detect individuals withBE.23-27,52-54 Although estimates varied among studies,endoscopic screening of individuals with GERD was found

www.giejournal.org

Page 9: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 5. Continued

No. of patients Effect

Certainty ImportanceSurveillance No surveillance/incomplete surveillance Relative (95% CI) Absolute (95% CI)

101/282 (35.8%) 144/249 (57.8%) RR .60 (.50-.71) 231 fewer per 1000(from 168 fewer to 289 fewer)

4BBB

VERY LOWCRITICAL

RR .73 (.57-.94) 4BBBVERY LOW

CRITICAL

71/209 (34.0%) 67/103 (65.0%) HR 1.27 (.78-2.07) 86 more per 1000(from 91 fewer to 236 more)

4BBBVERY LOW

CRITICAL

652/1256 (51.9%) 18906/23191 (81.5%) RR .48 (.37-.63) 424 fewer per 1000(from 302 fewer to 514 fewer)

4BBBVERY LOW

CRITICAL

HR .85 (.75-.95) 4BBB

VERY LOWCRITICAL

HR .85 (.75-.95) 4BBB

VERY LOWCRITICAL

383/686 (55.8%) 162/453 (35.8%) RR 2.11 (1.08-4.11) 397 more per 1000(from 29 more to 1000 more)

4BBB

VERY LOWCRITICAL

597/1180 (50.6%) 2631/23100 (11.4%) RR 5.52 (3.70-8.24) 515 more per 1000(from 308 more to 825 more)

4BBB

VERY LOWCRITICAL

AE: cardiopulmonary 1/170 to 1/10,000; perforation 1/2500 to 1/11,000

ASGE guideline on screening and surveillance of BE

to be cost-effective if the willingness-to-pay thresholdwas <$100,000 per quality-adjusted life-year.55

Discussion: The panel acknowledged the challengesassociated with this recommendation. Beyond the lack ofdata from RCTs, studies have shown that <10% of EAC pa-tients have a prior diagnosis of BE, indicating that currentpractices have failed to identify a large majority of pa-tients.56,57 Symptoms of GERD have been considered aprerequisite for screening for BE and EAC. However,only 7% to 10% of individuals with chronic GERD haveBE, and nearly 40% of EAC patients describe no prior his-tory of GERD.8,58 BE can be diagnosed in asymptomatic in-dividuals (those without any GERD symptoms), and nearly50% of patients with short-segment BE have no GERDsymptoms.48,59,60 These data suggest that a screening pro-gram that is limited to individuals with GERD symptomsmay miss a significant proportion of high-risk individuals.Furthermore, available data from EGDs performed at thetime of screening colonoscopy demonstrated a pooledprevalence of BE of more than 10%. The panel deliberated

www.giejournal.org

broadening the at-risk population to include individualswith multiple risk factors (age >50, male gender, whiterace, smoking, obesity) independent of GERD. However,this would drastically increase the number of patientseligible for BE screening and would incur increasedresource use and costs for patients, insurers, and thehealthcare system in general as well as increased potentialharm from endoscopy and patient burden. As other less-invasive and resource-intense modalities for BE screeningbecome available, this would influence future recommen-dations for screening.

Additionally, the panel recognized the value of risk pre-diction tools that use demographic and historical data.61-64

However, the accuracy of these prediction models remainsmodest with areas under the receiver operating character-istic curve of .61 to .75.65 Refinement and subsequentvalidation of these models will be critical beforewidespread implementation. Currently, standard upperendoscopy is the most common screening test used.However, other available modalities have been developed

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 343

Page 10: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 6. Association between various risk factors and BE based on literature review

Risk factor Study No. of studies Effect estimate

GERD Taylor et al48 26 OR Z 2.9 (4.5 in 12 studies with GERD for at least 2 weeks)

Family history of BE or EAC Chak et al58 1 OR Z 12, adjusted for age, sex, and obesity

Male Cook et al119 19 Male-to-female ratio Z 2.13

Obesity Singh et al50 10 OR Z 1.15 [.89-1.47], adjusting for GERD

Central adiposity Singh et al50 11 OR Z 1.44 [1.2-1.74], adjusting for GERD

Smoking Andrici et al49

Andrici et al49104

OR Z 1.42 [1.15-1.76], ever smoked vs population controlOR Z 1.96 [1.41-2.73], adjusted for confounders

OR, Odds ratio; BE, Barrett’s esophagus; EAC, Esophageal adenocarcinoma.

ASGE guideline on screening and surveillance of BE

with the goal of improving the effectiveness, reducing cost,and minimizing invasiveness of screening. Such modalitiesinclude transnasal endoscopy, esophageal capsule endoscopy,Cytosponge (Medtronic, Fridley, Minn), tethered capsuleendomicroscopy, and electronic nose device.56 In the SRand MA by Qumseya et al,51 standard upper endoscopy wasthe most common screening tool. Only 5 studies evaluatedother modalities: 3 studies66-68 used capsule endoscopyand 2 studies69,70 used ultrathin scopes. An SR and MA ofother novel screening technologies was not conducted forthis guideline document and is beyond the scope of this dis-cussion. A recent Clinical Practice Update document by theAmerican Gastroenterological Association reviewed the cur-rent status of these technologies and recommended againstthe use of any alternative test to screen for BE at this time.56

This panel also acknowledges the great potential andpromise for these technologies. After careful deliberationby all stakeholders, the panel endorsed screening for BE inat-risk populations for BE using the available evidence onthe prevalence of BE in specific populations.

Question 3: In patients with BE who are under-going surveillance for dysplasia, what is the roleof CE in increasing the rate of dysplasiadetection?

Recommendation: In patients with BE undergo-ing surveillance, we recommend using CE or VCin addition to WLE and biopsy specimens obtainedusing the Seattle protocol compared with WLE andbiopsy specimens obtained using the Seattle proto-col alone (strong recommendation, moderate qual-ity of evidence).

Summary of the evidence: The outcomes for this clin-ical question were the increase in diagnostic yield and perfor-mance characteristics of CE with WLE compared with WLEalone. We identified 2 existing SRs and MAs that addressedthis question. We updated a previously published SR andMA that addressed the clinical question of diagnostic yield us-ing CE.71 Additionally, an SR and MA performed by the ASGETechnology Committee was used to inform the questionaddressing performance characteristics of CE in BEsurveillance.72 This document assessed whether acceptable

344 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

performance thresholds outlined by the ASGE Preservationand Incorporation of Valuable Endoscopic Innovations(PIVI) document73 for clinical adoption of advancedimaging techniques have been met.

A new search was conducted using PubMed, Embase,and Web of Science to update the results from a previousSR and MA.71 We limited our search strategy to includeonly RCTs published in peer-reviewed journals. A total of1347 studies were identified, uploaded into Covidence.org, and reviewed by 2 independent reviewers. TwelveRCTs (n Z 2433) were identified (1 new study in additionto the 11 included in the previously published SR andMA).74 These studies were mostly crossover or tandem indesign. Using a random effects model, the absoluteincrease in dysplasia detection was 9% (95% CI, 4.1%-14%; P < .001; Q Z 29; I2 Z 42%). This analysis alsoaddressed the relative increase in dysplasia detection.Because the studies were crossover in design, we set theexternal correlation coefficient at .5. When a randomeffects model was used, the relative increase is dysplasiadetection was 30.3% (95% CI, 16.2%-44.3%; P < .0001).Varying the correlation coefficient had no significantdifference on the final results. In a subanalysis, there wasno significant difference between VC compared with dye-based CE (risk difference, 28.5% [95% CI, 12.3%-45.5%]vs 30.7% [95% CI, 5.2%-56.3%], P Z .98) (Fig. 1).

For the outcome of diagnostic accuracy, only 2 of theabove-mentioned RCTs reported the sensitivity and speci-ficity of CE.75,76 However, these were reported on a per-lesion analysis and not per-patient. The SR and MA pub-lished by the ASGE Technology Committee reported anoverall sensitivity of 91.9% (95% CI, 89.4%-93.8%), negativepredictive value of 95.5% (95% CI, 90.8%-97.9%), and spec-ificity of 89.9% (95% CI, 80.1%-95.2) for conventional CEusing dye-spraying.72 The diagnostic characteristics werehigher using acetic acid (sensitivity, 96.6% [95% CI, 95.2%-97.7%]; negative predictive value, 98.3% [95% CI, 94.8%-99.4%], and specificity, 84.6% [95% CI, 68.5%-93.2%])compared with methylene blue and met the thresholdsset by the ASGE PIVI document.73 VC using narrow-bandimaging (NBI) had an overall sensitivity of 94.2% (95%CI, 82.6%-98.2%), negative predictive value of 97.5%

www.giejournal.org

Page 11: ASGE guideline on screening and surveillance of Barrett's ...

Study name Statistics for each study Risk differenceand 95% CI

Riskdifference

Lowerlimit

Upperlimit P Value

Canto, M (2000)

Canto, M (2000)

Grossner L. (2006)

Grossner L. (2006)

Horwhat, J. (2007)

Horwhat, J. (2007)

Kara, M (2005)

Kara, M (2005)

Ragunath, K. (2003)

Ragunath, K. (2003)

Wo, J. (2001)

Wo, J. (2001)

Bartlie, S. (2015)

Bartlie, S. (2015)

Curvers, W. (2010)

Curvers, W. (2010)

Curvers, W. (2008)

Curvers, W. (2008)

Curvers, W. (2011)

Curvers, W. (2011)

Wolfsen, H. (2008)

Wolfsen, H. (2008)

Sharma, P. (2012)

Sharma, P. (2012)

Study name Std PairedDifference and 95% CI

0.163 -0.037

0.098

-0.233

-0.188

-0.129

-0.069

-0.018

0.038

0.000

-0.077

-0.032

-0.074

0.041

0.363

0.344

0.149

0.473

0.234

0.355

0.055

0.238

0.261

0.198

0.309

0.156

0.140

.111

.000

.669

.397

.570

.187

.312

.007

.049

.387

.111

.488

.000

-1.00 -0.50 0.00 0.50 1.00

-1.00 -0.50 0.00 0.50 1.00

Std PairedDifference

Lowerlimit

0.043

0.404-0427

-0.267-0.171

-0.029-0.3280.294

0.106-0.0830.037

-0.1060.162

0.7451.0240.226

0.4080.790

1.5611.0940.6650.357

0.578

0.3180.443

.028

.000

.546

.243

.422

.069.201

.001

.007

.223

.026

.328

.000

1.051

Upperlimit P Value

0.3940.714

-0.1010.392

0.1190.381

0.6160.694

0.3860.137

0.1060.303

0.307

0.221

0.1430.053

0.143

0.019

0.138

0.138

0.041

0.090

0.131

0.061

-0.042

Statistics for each study

A

B

Figure 1. Forest plots of the 12 clinical trials assessing the absolute (A) and relative (B) increase in the diagnostic yield of chromoendoscopy comparedwith white-light endoscopy in detection of dysplasia for patients with Barrett’s esophagus.

ASGE guideline on screening and surveillance of BE

(95% CI, 95.1%-98.7%), and specificity of 94.4% (95% CI,80.5%-98.6%), performance that exceeds the thresholdsset by the ASGE PIVIC document.73 We acknowledgethat per-patient analyses are more relevant and patient-centered and allow for a more accurate determinationof the impact of test findings on patient-importantoutcomes.

www.giejournal.org

Certainty in the evidence: For the critical outcome ofincrease in the diagnostic yield, 12 clinical trials wereincluded with high quality of evidence. There were no is-sues with inconsistency, imprecision, or publication bias(Fig. 2). However, because all studies were done atexpert centers, we rated down for indirectness, becauseapplicability to nonexpert centers may lead to different

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 345

Page 12: ASGE guideline on screening and surveillance of Barrett's ...

0.0

0.1

0.2

0.3

0.4

0.5

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0

Std Paired Difference

Stan

dard

Err

or

Funnel Plot of Standard Error by Std Paired Difference

Figure 2. Funnel plot assessing publication bias among studies on the diagnostic yield of chromoendoscopy and white-light endoscopy.

ASGE guideline on screening and surveillance of BE

estimates in diagnostic yield (Table 7). Therefore, thequality of evidence for the outcome of diagnostic yieldwas moderate.

Considerations: In making this recommendation, thepanel noted that the use of VC was essentially cost-equivalent: No extra cost is incurred because this technol-ogy is available in most endoscopes. Additionally, VC is alsorisk free to the patient. The panel also reviewed the evi-dence for dye-based CE, which appears to be cost-effective and can be used if VC is not available. Bhandariet al77 studied the cost-effectiveness of acetic acid inhigh-risk BE patients and found it to be more cost-effective than random biopsy sampling. Some dyes havebeen linked to potential risks, including DNA damage.78,79

Discussion: Several advanced imaging modalities havebeen investigated to improve the detection and identifica-tion of early neoplastic lesions during surveillance endos-copy.34 Dye-based CE refers to the application of varioustopical dyes or solutions to the esophageal mucosa forthe purpose of enhancing visibility of surface abnormal-ities. VC refers to using various light filters within the endo-scope to achieve similar visual enhancement. The 3available platforms for VC are NBI (Olympus, Center Valley,Pa), Fujinon intelligent color enhancement (Fujinon,Wayne, NJ), and i-Scan (Pentax Medical, Montvale, NJ).NBI is an imaging technique that is based on the opticalphenomenon that the depth of light penetration into tis-sue depends on the wavelength; the shorter the wave-length, the more superficial the penetration. Use of bluelight with narrow-band filters enables detailed imaging ofthe mucosal and vascular surface patterns with a highlevel of resolution and contrast without the need for

346 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

conventional CE.80 Fujinon intelligent color enhancementand i-Scan use proprietary postimage acquisitionprocessing technology to modify the white-light imageenhancing the superficial mucosal and vascular patterns.NBI is the most widely studied and used VC technique inclinical practice. An expert panel developed and validateda classification system to identify HGD and EAC in BE pa-tients using NBI.81

Based on our results, there is a strong body of evidenceto recommend this practice for all BE patients undergoingscreening/surveillance with VC as its first choice. The paneldid not make a specific recommendation for the type of VCto be used and acknowledged recent data demonstratingincreased dysplasia detection using technologies such asthe i-Scan Optical Enhancement system.82 For dye-basedCE, the panel made no specific recommendation for thetype of dye, but based on available data, acetic acid isthe only dye-based CE technique that meets the ASGEPIVI thresholds.73 Acetic acid enhances mucosal surfacepatterns by contrast staining. Problems associated withuse of dye-based CE in clinical practice include the needfor dye spraying equipment; difficulty in achieving com-plete and uniform coating of the mucosal surface withthe dye, inability to detect superficial vascular patterns,and the time-consuming and tedious nature of theprocedure.

Many have hoped the use of CE would help eliminatethe need for the random biopsy sampling. The processesof the Seattle protocol can be cumbersome and costly,especially in long-segment BE. The panel recognizes thatareas of BE may not appear to be suspicious on CE butmay still harbor dysplasia. Additionally, a meta-regression

www.giejournal.org

Page 13: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

by Qumseya et al71 showed that the number of biopsyspecimens strongly corresponded with the diagnosticyield. Therefore, the panel did not recommend CE as areplacement for the Seattle protocol but rather as anadjunct. The panel recognized that data from a singleRCT reported that NBI had similar detection of BE butrequired fewer biopsy specimens than high-definitionWLE in patients with BE undergoing screening or surveil-lance endoscopy.76

This panel also acknowledges that the studies reportedin this analysis include expert endoscopists recruiting pa-tients at tertiary care centers. There is a dearth of dataon the learning curves and impact of training in the detec-tion of early neoplasia (with or without advanced imagingtechniques). The future lies in incorporating training in CEin gastroenterology fellowship programs and establishingtraining programs for detection of early neoplasia thatare acceptable and applicable to the broader gastroenter-ology community.

Question 4: In patients with BE undergoingendoscopy for surveillance of dysplasia, what isthe role of CLE in increasing the rate of dysplasiadetection?

Recommendation: In patients with BE undergo-ing surveillance, we suggest against routine useof CLE compared with WLE with Seattle protocol bi-opsy sampling (conditional recommendation, lowquality of evidence).

Summary of the evidence: The outcomes of interestfor this question were the increase in diagnostic yield ofCLE (critical) and performance characteristics of CLE(important). We identified 2 existing SRs: the previouslydescribed SR and MA by the ASGE Technology Commit-tee72 and another study by Xiong et al.83 We updated themore recent SR by Xiong et al and used this to addressour clinical question. An updated search from June 2015until June 2017 identified 205 new studies, and none waseligible for inclusion. Therefore, four RCTs detailing thediagnostic yield in CLE versus WLE were included.84-87

Two studies reported using probe-based CLE, whereasthe other 2 used endoscope-based CLE. Using a random ef-fects model and a per-patient analysis, the absolute in-crease in dysplasia detection using CLE was 10.2% (95%CI, 1.4%-19.1%; P Z .024; I2 Z 42%). Note that the CIfor this measure was very close to zero. The relative in-crease in dysplasia detection was not significant: 36%(95% CI, –5.4% to 77.5%; P Z .088; I2 Z 64%) (Fig. 3,Table 8).

For performance characteristics (sensitivity and speci-ficity), Xiong et al83 analyzed 7 studies totaling 473patients and reported a pooled sensitivity of 89% (95%CI, 82%-94%) and a pooled specificity of 83% (95% CI,78%-86%). The overall prevalence of dysplasia/neoplasiavaried significantly by study ranging from 12.5% to 100%.

www.giejournal.org

Table 9 highlights the performance of CLE varying theprevalence of dysplasia (5% and 30%) in the underlyingpopulation.

The panel also considered AEs associated with CLE. Themain risks of CLE relate to the need to inject intravenousfluorescein to conduct these procedures. Wallace et al88

conducted a cross-sectional international survey of 16 aca-demic centers. Based on 2272 CLE procedures, theyfound no major AEs. The rate of minor adverse outcomeswas 1.4%, and this included epigastric pain, nausea/vomiting,rash, injection site erythema, and transient hypotension.

Certainty in the evidence: The certainty in the evi-dence for the diagnostic yield was based on data fromthe 4 RCTs. We rated down for inconsistency given thehigh I2 and for indirectness given that all studies weredone at tertiary referral centers. Therefore, the overallquality of evidence was rated as low (Table 8). For theoutcome of sensitivity and specificity, we rated down forimprecision given the wide CIs and inconsistency onaccount of the high I2. Therefore, the quality of evidencewas also rated as low (Table 9).

Considerations: The panel considered the AEs for CLEas described above.88 The panel also considered that thistechnology requires up-front investment in equipmentand significant training to achieve proficiency.89 Therewere no studies regarding patient values and preferencesusing this technology, and the overall cost-effectivenessof CLE has not been adequately studied. A cost-utility anal-ysis using Markov modeling showed that the routine use ofCLE and optical coherence tomography was not cost-effective compared with the use of WLE with the Seattleprotocol biopsy sampling.90

Discussion: CLE uses blue laser light to illuminate is-sues after application of intravenous fluorescence agents.This can produce magnification up to 1250-fold, allowingvisualization at the cellular level, with the resulting abilityto identify areas of dysplasia/neoplasia. For CLE, as withCE, the panel placed a higher level of importance on theincrease in the diagnostic yield compared with the perfor-mance characteristics because of reasons discussed earlier.Our results showed that the absolute increase in the diag-nostic yield for CLE was 10%, similar to that reported forCE. However, the CI was very close to zero. Additionally,the relative increase in the diagnostic yield was 36%,but this was not statistically significant (CI crossed zero).The panel assumed that a minimum of 20% relative in-crease in diagnostic yield would be an appropriatethreshold that might warrant a recommendation forCLE. This threshold was partly based on a recent studyamong a group of international BE experts who reportedthe minimum incremental diagnostic yield for the use ofadvanced imaging techniques in BE surveillance.91 In thisstudy, experts reported a minimum incremental increasein the diagnostic yield of 27% (interquartile range, 20%-50%) for CLE before implementation in clinical practice.

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 347

Page 14: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 7. Evidence profile for use of chromoendoscopy compared with white-light endoscopy for detection of dysplasia in patients with Barrett’sesophagus

Certainty assessment

No. of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations

Chromoendoscopy: increased yield in dysplasia detection

12 Randomized trials Not serious Not serious Serious* Not serious None

RR, Relative risk.*Most studies were done at tertiary centers with gastroenterologists who have expertise and skills in detecting Barrett’s esophagus, and these findings may not be generalizableto all gastroenterologists

TABLE 8. Evidence profile for use of confocal laser endomicroscopy compared with white-light endoscopy for detection of dysplasia in patientswith Barrett’s esophagus

Certainty assessment No. of patients

No. of studies Study design Risk of bias Inconsistency Indirectness ImprecisionOther

considerationsConfocal laser

endomicroscopyWhite-lightendoscopy

Increased yield in dysplasia detection

4 Randomized trials Not serious Serious* Seriousy Not serious None 75/284 (26.4%) 49/288 (17.0%)

RR, Relative risk.*We rated down for inconsistency.yStudies were done at tertiary centers with gastroenterologists who have expertise and skills in detecting Barrett’s esophagus, and these findings may not be generalizable to allgastroenterologists.

A

B

Study name

Study name Comparison

Std PairedDifference

Std Paired Differenceand 95% CI

P Value

Bertano, 2013

Sharma, 2011

Dunbar, 2009

Canto, 2014

Bertano, 2013

Sharma, 2011

Dunbar, 2009

Canto, 2014

pCLE

pCLE

eCLE

eCLE

0.691 0.078 1.303

1.000

1.346

0.775

.027

.753

.002

.088

0.394

0.544

0.290

-0.285

-0.054

-0.544

0.055

0.000

0.818

0.360

Statistics for each study

Statistics for each study

Riskdifference

Risk differenceand 95% CI

-1.00 -0.50 0.00 0.50 1.00

Lowerlimit

Lowerlimit

Upperlimit

Upperlimit

P Value

0.180 0.030 0.330 .018

.753

.001

.024

1.000

0.143

0.200

0.259

0.191

-0.104

-0.200

-0.066

-0.014

0.020

0.000

0.162

0.102

Figure 3. Forest plots of the 4 clinical trials assessing the absolute (A) and relative (B) increase in the diagnostic yield of confocal laser endomicroscopycompared with white-light endoscopy in detection of dysplasia for patients with Barrett’s esophagus. pCLE, probe-based confocal laser endomicroscopy;eCLE, endoscope-based confocal laser endomicroscopy.

348 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019 www.giejournal.org

ASGE guideline on screening and surveillance of BE

Page 15: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 7. Continued

No. of patients Effect

Certainty ImportanceChromoendoscopy White-light endoscopy Relative (95% CI) Absolute (95% CI)

408/848 (48.1%) 329/848 (38.8%) RR 1.31 (1.21-1.41) 120 more per 1000(from 81 more to 159 more)

444B

MODERATECRITICAL

TABLE 8. Continued

Effect

Certainty ImportanceRelative (95% CI) Absolute (95% CI)

RR 1.50 (.93-2.08) 85 more per 1000(from 12 fewer to 184 more)

44BBLOW

CRITICAL

ASGE guideline on screening and surveillance of BE

Furthermore, the 4 RCTs had a prevalence of dysplasiaranging from 19% to 35%, much higher than theprevalence of dysplasia in the BE population seen by mostgastroenterologists at nonspecialized centers. Therefore,in practice, CLE is likely to have a lower diagnostic yield.Therefore, our results indicate that the benefit from CLEis small at best in the general surveillance population. Theadded cost of equipment, training, and possible risk ofintravenous dye administration makes the routine use ofCLE less desirable.

With regard to the diagnostic characteristics, there arelimited data on patient-related outcomes. For example,there are no data on how many patients who had a nega-tive CLE subsequently developed dysplasia. The SR andMA by the ASGE Technology Committee reported an over-all (combining endoscope-based CLE and probe-basedCLE) sensitivity of 90.4% (95% CI, 75.7%-96.6%), negativepredictive value 96.2% (95% CI, 93.1%-97.9%), and speci-ficity of 89.9% (95% CI, 83.8%-93.9%), thresholds that didnot meet the ASGE PIVI thresholds.72 Althoughendoscope-based CLE met the ASGE PIVI thresholds, thiswas based on data from studies conducted at tertiarycare centers. In addition, this technology (endoscope-based CLE) is no longer available in clinical practice. Asillustrated in the evidence profile (Table 9), in a low-prevalence population, the false-positive rate was relativelyhigh at 16.3% (95% CI, 13.4%-21.1%), but even with a prev-alence of 30%, the false-positive rate remains high at 12%.Although this guideline suggests against the routine use ofCLE in BE patients, we acknowledge that CLE may be ahelpful tool in increasing the diagnostic yield of dysplasiain centers with a high prevalence of dysplasia and signifi-cant local expertise.

www.giejournal.org

Question 5: What is the role of EUS in staging BEpatients with dysplasia or early EAC?

Recommendation: In BE patients with dysplasiaor early EAC, we recommend against routine useof EUS to differentiate mucosal versus submucosaldisease (strong recommendation, moderate qual-ity of evidence).

Summary of the evidence: This guideline statementpertains to the use of EUS in differentiating mucosal(T1a) from submucosal (T1b) EAC. The rationale for usingEUS in this patient population is to detect advanced dis-ease into and beyond the submucosa with the goal ofavoiding potentially invasive EMR or endoscopic submuco-sal dissection in patients with advanced disease and avoid-ing esophagectomy in patients with EAC limited to themucosa (T1a disease). The patient-important outcomesin this analysis were mortality and cancer progressionand the diagnostic performance of EUS to detect advanceddisease. We identified no studies that addressed theoutcome of mortality and cancer progression. For theoutcome of diagnostic characteristics of EUS in BE patientswith dysplasia and early EAC, we started with 2 MAs byQumseya et al.92,93 These analyses focused on the perfor-mance characteristics of EUS at the T1a versus T1b level us-ing data from 11 studies. The pooled sensitivity, specificity,and accuracy from this analysis were 41% (95% CI, 35%-48%), 89% (95% CI, 86%-91%), and 75% (95% CI, 59%-86%), respectively. As with other diagnostic studies, we re-viewed performance characteristics based on the preva-lence of disease. In the included studies, the prevalenceof advanced disease ranged from 5% to 45%. The evidenceprofile illustrates how the false-positive rate increases from6% in patients with high prevalence of advanced disease to

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 349

Page 16: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 9. Evidence profile for performance characteristics of confocal laser endomicroscopy compared with white-light endoscopy for detectionof dysplasia in patients with Barrett’s esophagus using a prevalence of 5% and 30%*

Sensitivity .89 (95% CI, .82-.94) Prevalences 5% 30%

Specificity .83 (95% CI, .78-.86)

Outcome

No. ofstudies(no. ofpatients)

Studydesign

Factors that may decrease certainty of evidenceEffect per 1000 patients

tested

Testaccuracy

Riskofbias Indirectness Inconsistency Imprecision

Publicationbias

Pretestprobability

of 5%

Pretestprobabilityof 30%

True positives(patients withdysplasia)

7 studiespatients

Cross-sectional

(cohort typeaccuracy study)

Notserious

Not serious Seriousy Seriousz None 45(41-47)

267(246-282)

44BBLOW

False negatives(incorrectlyclassified asnot havingdysplasia)

5 (3-9) 33(18-54)

True negatives(patientswithoutdysplasia)

7 studiespatients

Cross-sectional

(cohort typeaccuracy study)

Notserious

Not serious Seriousy Seriousz None 789(741-817)

581(546-602)

44BB

LOW

False positives(incorrectlyclassified ashavingdysplasia)

161(133-209)

119 (98-154)

*The illustrated prevalences of 5% and 30% were used as estimates of the prevalence of Barrett’s esophagus and/or dysplasia in patients referred for EGD in the community andat tertiary care centers. These estimates were based on expert opinion and the best available published evidence.yWe rated down for inconsistency, as there was significant heterogeneity.zWe rated down for imprecision.

ASGE guideline on screening and surveillance of BE

10.5% in patients with low prevalence of advanced disease(Table 10). Of the 11 studies in the MA, the prevalence ofHGD/EAC was 100% in all but 2 studies, where theprevalence of dysplasia was 49% and 65%.94,95 A prior MAby Yousef et al96 reported a crude HGD/EAC prevalencerate of 4% (178 patients with HGD/EAC among 4491 BEpatients).

Certainty in the evidence: We rated down for incon-sistency given heterogeneity. There was no evidence ofpublication bias, inconsistency, or indirectness. Therefore,the quality of evidence was rated as moderate (Table 10).

Considerations: There are no studies assessing thecost-effectiveness of EUS for this patient population. Simi-larly, there are no direct data on AEs from EUS in this pa-tient group. However, based on indirect evidence, EUS isassociated with a risk of cervical esophageal perforationestimated at .03% to .06%97 and mortality of .002%.98

Bleeding after EUS with FNA has been reported in .13%based on a MA.99 Finally, rare cases of tumor seedinghave been reported.100 There were no reports on patientvalues or preferences other than the input we receivedfrom the patient representative on our panel.

Discussion: The use of EUS in BE continues to becontroversial, although there seems to be a trend toward do-ing less EUS in this patient population because of the lack ofreliability of EUS in accurately distinguishing between pa-

350 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

tients with HGD/T1a cancer from those with T1b cancer.Our evidence profiles showed that EUS is associated with arelatively high false-positive rate ranging from 6% to 10%.

Of all patients who test positive for advanced disease onEUS, around one third will be false positive for advanced dis-ease. In light of the possible AEs of EUS, potential costs ofprocedure and sedation, and the high false-positive rate,the panel decided to recommend against the routine useof EUS for this specific indication. It should also be empha-sized that the most recent ASGE guideline document on EETstrongly recommends endoscopic resection of all visible le-sions in BE patients as first-line modality for diagnostic andtherapeutic purposes.2 This recommendation is in linewith other GI society guidelines and quality indicatordocuments.4,101 This recommendation should not beconfused with the role of EUS in patients with EAC. EUS isindicated in patients with EAC for accurate staging ofadvanced cancer (�T1b) and to evaluate for nodal disease.

A study that used the Surveillance, Epidemiology,and End Results Medicare database showed that thereceipt of EUS was a significant predictor of improved1-year (HR, .49; 95% CI, .39-.59; P < .0001), 3-year(HR, .57; 95% CI, .48-.66; P < .0001), and 5-yearsurvival (HR, .59; 95% CI, .5-.68) driven primarily by pro-vision of stage-appropriate treatment for EAC pa-tients.102 Receipt of EUS also increased the likelihood

www.giejournal.org

Page 17: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 10. Evidence profile for performance characteristics of EUS compared with no EUS for diagnosing advanced neoplasia using a prevalenceof 5% and 30%*

Sensitivity .41 (95% CI, .35-.48) Prevalences 5% 30%

Specificity .89 (95% CI, .86-.91)

Outcome

No. ofstudies(no. ofpatients)

Studydesign

Factors that may decrease certainty of evidenceEffect per 1000 patients

tested

Testaccuracy

Riskofbias Indirectness Inconsistency Imprecision

Publicationbias

Pretestprobability

of 5%

Pretestprobabilityof 30%

True positives(patients withadvancedneoplasia)

11 studies228 patients

Cross-sectional(cohorttype

accuracystudy)

Notserious

Not serious Seriousy Not serious None 21(17-24)

123(105-144)

444BMODERATE

False negatives(patientsincorrectlyclassified asnot havingadvancedneoplasia)

29(26-33)

177(156-195)

True negatives(patientswithoutadvancedneoplasia)

11 studies679 patients

Cross-sectional(cohorttype

accuracystudy)

Notserious

Not serious Seriousy Not serious None 845(817-864)

623(602-637)

444BMODERATE

advancedneoplasia)

105(86-133)

77(63-98)

*The illustrated prevalence of 5% and 30% were used as estimates of the prevalence of Barrett’s esophagus and/or dysplasia in patients referred for EGD in the community andat tertiary care centers. These estimates were based on expert opinion and the best available published evidence.yWe rated down for inconsistency.

ASGE guideline on screening and surveillance of BE

of receiving endoscopic therapies, esophagectomy, andchemoradiation. Therefore, the panel acknowledgedseveral circumstances in which a clinician may considerperforming EUS. However, EUS should not be used forT staging of early disease instead of EMR or endoscopicsubmucosal dissection of such lesions.

Question 6a: In patients with known or sus-pected BE, what is the role of WATS withcomputer-assisted 3-dimensional analysis (WATS-3D) in increasing the rate of dysplasia detection?

Recommendation: In patients with known or sus-pected BE, we suggest using WATS-3D in addition toWLE with Seattle protocol biopsy samplingcompared with WLE with Seattle protocol biopsysampling alone (conditional recommendation,low quality of evidence).

Summary of the evidence: A new SR was conductedto address this clinical question. The search strategy re-sulted in 1554 studies (search date December 15, 2018).After review by 2 independent reviewers, 40 studies werereviewed in detail and 6 studies (5 trials and 1 meeting ab-stract) were included in the final analysis.

Of the 6 studies, 3 studies102-104 were in patients withBE and dysplasia (high-risk group) who were undergoingsurveillance and 3 studies105-107 included patients with

www.giejournal.org

and without a history of dysplasia. Patients without a his-tory of BE were removed from all analyses. Therefore,6271 patients with BE were analyzed. WLE with random bi-opsy sampling detected 125 cases of dysplasia. The perfor-mance of WATS resulted in identification of 137 additionalcases missed by WLE with random biopsy sampling.Using a random effects model, the relative increase indysplasia detection was 48% (95% CI, 34%-60%%; I2 Z68%, Q Z 16). All heterogeneity in this analysis resultedfrom 1 study by Smith et al.105 Removing the studyreduced I2 to zero.

For patients with history of dysplasia, the relative in-crease in dysplasia detection using WATS was 47%(95% CI, 32%-61%). The absolute increase in dysplasiadetection using WATS was 10.6% (95% CI, 1.5%-19.8%).The relative increase in LGD detection was 21% (95%CI, 24%-40%) (Table 11, Fig. 4). For studies reportingall patients with or without a history of dysplasia,referred to as all-comers, the relative increase indysplasia detection was 52% (95% CI, 21%-82%), whereasthe absolute increase in dysplasia detection was 2% (95%CI, 1.5%-2.5%) (Table 11, Fig. 4). Four of the 6 studiesreported the yield of WATS in detection of LGD. Whenthe random effects model was used, the absoluteincrease in LGD detection was 1.8% (95% CI, 1.4%-2.3%) (Fig. 5).

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 351

Page 18: ASGE guideline on screening and surveillance of Barrett's ...

TABLE 11. Evidence profile for use of wide-area transepithelial sampling for detection of dysplasia in patients with Barrett’s esophagus

Certainty assessment

No. of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations

Dysplasia detection

6 Randomized trials Serious* Seriousy Not serious Not serious None

Adverse events

1 Observational studies Seriousz Not serious Not serious Not serious None

WATS-3D, Wide-area transepithelial sampling with computer-assisted 3-dimensional analysis; WLEþRB, white-light endoscopy with random biopsies; RR, relative increase indysplasia detection.*Dysplasia definition varied.yModerate heterogeneity in some of the analyses.zOnly 1 study in abstract form.

ASGE guideline on screening and surveillance of BE

Certainty in the evidence: We rated down for risk ofbias, inconsistency, and indirectness. The dysplasia defini-tion varied across studies; some studies included LGD,whereas others did not. Additionally, most studies werefunded by the manufacturer. There was no evidence ofpublication bias (Fig. 6). Therefore, the quality ofevidence was low (Table 11).

Considerations: To date, we found no studiesassessing cost-effectiveness of WATS-3D in routine sur-veillance of BE patients. With respect to AEs, wecontacted the authors who reported no significantadverse outcomes from using this technology. We alsoidentified 1 abstract by Smith et al108 that surveyed 33physicians using WATS-3D with over 4881 cases ofWATS-3D. They reported rate of serious adverse out-comes at .06%. Patient preferences have not been re-ported to date.

Discussion: The use of WATS-3D has received muchattention in recent years. The idea of WATS-3D is novelin that although previous technologies have focused onattempting to increase dysplasia detection by improvingvisualization of dysplastic areas, WATS-3D attempts toimprove dysplasia detection by increasing the surfacearea sampled. Unlike standard soft cytology brushes, theWATS-3D biopsy instrument samples deeper layers of themore firmly attached glandular epithelium in the esoph-agus. Unlike histology specimens from forceps biopsy sam-pling that are sliced to a 1- to 3-mm thickness, samplesusing this technology are uncut to preserve the 3D repre-sentation of the intact cellular structure. The analysis of thesample is aided by a high-speed computer scan, whichidentifies potentially abnormal cells, cell clusters, andabnormal glandular cells on a high-resolution videomonitor for pathology review. The most suspicious cellsare flagged by the computer as a starting point for the pa-thologists to review.

An important limitation of the above studies was howdysplasia was defined. All the aforementioned studieswere funded by themanufacturer (CDxDiagnostics, Suffern,NY). In an attempt to stratify data by dysplasia type, the

352 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

manufacturer was contacted by the lead author (B.Q.) andthe chair of the Standards of Practice committee (S.W.).For Anandasabapathy et al,102 Johanson et al,107 and Iorioet al,104 the outcome reported was dysplasia. Themanufacturer confirmed that indefinite for dysplasia wasexcluded from the analyses in 2 of the 3 studies. In Ioriet al, 3 of the 4 cases detected by WATS were LGD and 1was HGD. In the study by Smith et al,108 WATS detected10 cases of HGD/EAC and 75 cases with LGD. For ouranalyses, the cases of indeterminate for dysplasia wereremoved. Therefore, we note that the increase in dysplasiadetection reported on WATS-3D may be largely related toLGD. In a subanalysis of 4 studies, the absolute increase inLGD detection was 1.8%. However, the question remainsof what LGD means on cytology compared with LGDdysplasia on histology. The minimum incremental diag-nostic yield of dysplasia for the use of WATS-3D, based ona recent survey of an international group of BE experts,91

was 27% (95% CI, 20%-50%). Results from our SR and MAdemonstrate that WATS-3D exceeds this threshold.

Another related question is the diagnosis of cryptdysplasia reported on WATS specimens. This is diagnosedwhen the pathologist reports dysplasia-like changes thatseem to be isolated to the crypts and not to the surfaceepithelium. Shaheen et al109 recently presented data ondisease progression in patients with crypt dysplasia (n Z310) and noted a progression rate of 2.1% per patient yearfor an endpoint of HGD/EAC. In the same study, the rateof disease progression in 83 patients with LGD was 7.7%per patient-year. This rate of 2.1% reported by Shaheenet al is very similar to the incidence rate of 2.2% reportedin a MA by Qumseya et al11 assessing disease progressionrates in BE patients with LGD. This suggests that patientswith the diagnosis of crypt dysplasia may be at a higherrisk for progression and future prospective studies shoulddefine the natural history of this entity beforemanagement plans are altered based on this diagnosis.

The panel initially made no recommendation for WATS-3D at the face-to-face meeting. After a complete review ofadditional published literature (including data on AEs) and

www.giejournal.org

Page 19: ASGE guideline on screening and surveillance of Barrett's ...

Riskdifference WATS WLE+RB

Relativeweight

All corners

All corners

All corners

Subgroup within study

All corners

All corners

All corners

Gross 2017

Smith 2017

History of Dysplasia

History of Dysplasia

History of Dysplasia

History of Dysplasia

History of Dysplasia

History of Dysplasia

0.021

0.020

0.020

0.093

0.206

0.036

0.106

0.021

49 / 1087 26 / 1087

26 / 4254

16 / 509

10.40

85.83

34.37

29.79

35.84

0.36

0.81

0.36

0.54

0.42

0.54

0.27

0.47

0.51

Study name

Study name

dysplasia / Total

Dysplasia / Total

Std PairedDifference WATS+biopsies Biopsies

3.76

111 / 4254

30 / 509

29 / 151 15 / 151

35 / 16068 / 160

11 / 110 7 / 110

49 / 1087 26 / 1087

26 / 4254

16 / 509

111 / 4254

30 / 509

29 / 151 15 / 151

35 / 16068 / 160

11 / 110 7 / 110

0.028

Vennalaganti 2017

Iorio 2015

Johnason 2011

Gross 2017

Smith 2017

Johnason 2011

Anandasbapathy 2011

Vennalaganti 2017

Iorio 2015

Anandasbapathy 2011

A

B

-0.25-0.130.00 0.13 0.25

-1.000.500.000.501.00

Figure 4. Forest plots of the 6 studies assessing the absolute (A) and relative (B) increase in the diagnostic yield of wide-area transepithelial samplingcompared with white-light endoscopy in detection of dysplasia for patients with Barrett’s esophagus. WATS, Wide-area transepithelial sampling; WLEþRB,white-light endoscopy with random biopsies.

TABLE 11. Continued

No. of patients Effect

Certainty ImportanceWATS-3D WLE D RB Relative (95% CI) Absolute (95% CI)

137/6271 (2.2%) 125/6271 (2.0%) RR 2.25 (1.79-2.83) 25 more per 1000 (from 16 more to 36 more) 44BB

LOWCRITICAL

Rate of serious adverse events .06% in survey of 33 physicians 4BBB

VERY LOWCRITICAL

www.giejournal.org Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 353

ASGE guideline on screening and surveillance of BE

Page 20: ASGE guideline on screening and surveillance of Barrett's ...

A

B

Study name

Study name Statistics for each study

Subgroup within study

Subgroup within studyGroup by

Dysplasia / Total

Risk

difference WATS

-0.17 -0.09 0.00 0.08 0.17

WLE+RBRelative

weight

Relativeweight

50.51

49.49

77.26

22.74

Std PairedDifference Variance

Lowerlimit P Value

0.36

0.92

0.64

0.21

0.01 0.17 0.55 .00

.00

-1.00 -0.50 0.00 0.50 1.00

.00

.02

.05

.03

.29

1.14

1.18

0.43

0.61

0.40

0.43

0.69

0.09

-0.00

-0.18

0.02

0.08

0.04

0.01

0.01

0.01

0.01

0.08

0.21

0.21

0.26

Upperlimit

Gross 2017

Smith 2017

Gross 2017

Smith 2017

Vennalaganti 2017

loria 2015

Vennalaganti 2017

loria 2015

All corners

All corners

All corners

All corners

All corners

History of Dysplasia

History of Dysplasia

History of Dysplasia

History of Dysplasia

Overall

History of Dysplasia

0.021

0.018

0.063

0.027

0.018

8.91

90.41

0.27

0.42

49 / 1087 26 / 1087

18 / 4254

28 / 160

93 / 4254

38 / 160

10 / 110 7 / 110

Figure 5. Forest plots of the assessing the absolute (A) and relative (B) increase in the diagnostic yield of wide-area transepithelial sampling comparedwith white-light endoscopy in detection of low-grade dysplasia for patients with Barrett’s esophagus. WATS, Wide-area transepithelial sampling; WLEþRB,white-light endoscopy with random biopsies.

ASGE guideline on screening and surveillance of BE

an additional phone conference, the panel made a condi-tional recommendation for the use of WATS-3D in additionto WLE with Seattle protocol biopsy sampling.

Question 6b: In patients with BE undergoingendoscopy for surveillance of dysplasia, what isthe role of VLE in increasing the rate of dysplasiadetection?

Recommendation: There is insufficient evidenceto make a recommendation for or against routineuse of VLE in surveillance of patients with BE (norecommendation).

Summary of the evidence: For this clinical question,we started with an existing SR and MA currently publishedin abstract form.110 This study included a search ofMedline, Embase, Web of Science, and Cochrane Centralending in October 2016. Of 487 studies, only 4 abstractswere identified. When the random effects model wasused, the pooled absolute increase in dysplasia detectionwas 2.3% (95% CI, 1.5%-3.4%) with a false-positive rate of44%. However, this study was only in abstract form andthe 4 studies included stemmed from the same database,and therefore we could not rely on this study for our panel.The main issue with CLE is that the technology has

354 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

continued to evolve. Most recently, the technology now in-cludes the ability to use laser tagging that allows bettermarking of areas of suspected dysplasia identified byVLE. After tagging such areas, biopsy sampling can bedone to confirm the presence or absence of dysplasia.Henceforth, the evidence on this technology is stillevolving. Based on our updated search and after contactingthe manufacturer, only 1 study111 was identified thatassessed the increase in dysplasia detection using thenew VLE laser device. This study by Alshelleh et al111

compared results from biopsy specimens obtained usingthe Seattle protocol, random biopsy sampling, VLE, andVLE laser (with laser marking). This study used anhistorical analysis comparing these techniques. Theauthors analyzed 386 patients. Overall, the studyreported higher dysplasia detection in the VLE lasergroup (33.7%) compared with the Seattle protocol group(19.6%). In post-treatment surveillance, total dysplasiadetection was higher in VLE laser compared with the Seat-tle biopsy sampling group (8.3% vs 32.7%, P Z .02).Finally, in treatment-naïve patients undergoing surveil-lance, the rate of dysplasia detection was not different inVLE laser compared with the Seattle protocol group(16.7% vs 35.3%) (Table 12).

www.giejournal.org

Page 21: ASGE guideline on screening and surveillance of Barrett's ...

0.00

0.05

0.10

0.15

0.20

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0

Std Paired Difference

Stan

dard

Err

or

Funnel Plot of Standard Error by Std Paired Difference

Figure 6. Funnel plots showing risk of publication bias for studies assessing wide-area transepithelial sampling in patients with Barrett’s esophagus.

TABLE 12. Overall dysplasia detection as presented by Alshelleh et al111

Group RB (n [ 79) SP (n [ 95)Volumetric laser

endomicroscopy (n [ 168)Volumetric laser

endomicroscopy laser (n [ 106)

Years 2011-2015 2011-2015 2014-2016 2016-2017

Overall dysplasia detection, % 5.7 19.6 24.8 33.7

RB, Random biopsies; SP, Seattle protocol.

ASGE guideline on screening and surveillance of BE

Certainty in the evidence: Because there was only 1cohort study and limited data available in abstract form,the overall quality of evidence was very low.

Considerations: There are no data on patient prefer-ences or cost-effectiveness of VLE in routine surveillanceof BE patients. Similarly, detailed reports on AEs are un-available. However, the panel noted that the proceduredoes include balloon inflation, which has the potential tocause pain or even perforation. The procedure does alsorequire additional time to introduce the device, inflatethe balloon, and obtain and interpret images. The exactcost of such measures is not clear at this time.

Discussion: VLE is based on optical coherence tomog-raphy and is an advanced imaging technique that producesa complete scan of the esophageal wall, including subsur-face layers, with a resolution comparable with low-powermicroscopy.112 Optical coherence tomography uses lightwaves instead of sound waves to form 2-dimensional im-ages based on differences in optical scattering of tissuestructures. With second-generation optical coherence to-mography, it is possible to perform high-resolution, high-speed acquisition of large luminal surfaces and ultimatelycreating 3D imaging of the esophagus wall.113 VLE scans6 cm of the esophagus over 90 seconds and provides aresolution of 10 mm and an imaging depth of 3 mm.

The use of VLE, especially VLE laser, is an emerging area.The panel acknowledged the limitations in the current

www.giejournal.org

studies. In light of evidence gaps, the panel concludedthere was insufficient evidence to make a recommendationfor this question. In addition to the thresholds set by theASGE PIVI document, international experts in BE reporteda minimum incremental increase in the diagnostic yield ofdysplasia of 30% (95% CI, 18%-50%) for the use of VLE inroutine clinical practice.91 Validation of VLE features andthe VLE prediction score for neoplasia and demonstratingthat these thresholds can be achieved in large prospectivetrials is required.114,115 Finally, studies defining the role ofcomputer-aided analysis as an aid in VLE interpretationare awaited.116

FUTURE DIRECTIONS

This document highlights several knowledge gaps in thefield of screening and surveillance for BE. First and fore-most, we still lack high-quality evidence on the benefitsof screening and surveillance in patients with BE specif-ically addressing key outcomes such as incidence,morbidity, and mortality associated with EAC. In this docu-ment, we discuss the best design for future studiesto address this critical issue. Future studies that refineand validate existing prediction tools for screening ofBE and EAC are required. These tools may require theaddition of noninvasive genetic or blood biomarkers to

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 355

Page 22: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

demographic and historical variables to improve the areasunder the receiver operating characteristic curves andoverall performance.117 Before we embrace the newgeneration of less-invasive and potentially less-expensivescreening techniques and replace our current approachof using standard endoscopy for screening, these newtechniques need to demonstrate high diagnostic perfor-mance characteristics, easy implementation at a primarycare level, high uptake in the at-risk population, and lowcost.118 Future studies also need to focus on improvedrisk stratification of BE patients undergoing surveillancewith the intent of performing EET for high-risk populationsand extending or discontinuing surveillance in low-riskgroups. Finally, effectiveness and validation data regardingadvanced imaging and sampling techniques among nonex-pert endoscopists are needed.

SUMMARY AND CONCLUSIONS

In this document, the ASGE offers evidence-based clin-ical practice guidelines on topics regarding screening andsurveillance for BE. These guidelines follow the GRADEframework and offer guidance on several key clinical ques-tions such as the role and impact of screening and surveil-lance in patients with BE and the role of advanced imagingtechniques in BE patients undergoing surveillance endos-copy. This guideline complies with the standards for guide-line development set forth by the Institute of Medicine forthe creation of trustworthy guidelines and aims to help cli-nicians understand the published literature and the qualityof available data with the ultimate goal of optimizing carefor patients with BE.

ACKNOWLEDGMENT

The panel thanks John Bastian for his contributionsfrom the patient perspective.

REFERENCES

1. Wani S, Sultan S, Qumseya B, et al. The ASGE’S vision for developingclinical practice guidelines: the path forward. Gastrointest Endosc2018;87:932-3.

2. Standards of Practice Committee; Wani S, Qumseya B, Shahnaz S,et al. Endoscopic eradication therapy for patients with Barrett’sesophagus-associated dysplasia and intramucosal cancer. Gastroint-est Endosc 2018;87:907-31.

3. Wani S, Rubenstein JH, Vieth M, et al. Diagnosis and management oflow-grade dysplasia in Barrett’s esophagus: expert review from theClinical Practice Updates Committee of the American Gastroentero-logical Association. Gastroenterology 2016;151:822-35.

4. Shaheen NJ, Falk GW, Iyer PG, et al. ACG clinical guideline: diagnosisand management of Barrett’s esophagus. Am J Gastroenterol2016;111:30-50; quiz 51.

5. Thrift AP. Barrett’s esophagus and esophageal adenocarcinoma: Howcommon are they really? Dig Dis Sci 2018;63:1988-96.

356 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

6. Rubenstein JH, Shaheen NJ. Epidemiology, diagnosis, and manage-ment of esophageal adenocarcinoma. Gastroenterology 2015;149:302-17.

7. Hur C, Miller M, Kong CY, et al. Trends in esophageal adenocarcinomaincidence and mortality. Cancer 2013;119:1149-58.

8. Ronkainen J, Aro P, Storskrubb T, et al. Prevalence of Barrett’s esoph-agus in the general population: an endoscopic study. Gastroenter-ology 2005;129:1825-31.

9. Shaheen NJ, Sharma P, Overholt BF, et al. Radiofrequency ablation inBarrett’s esophagus with dysplasia. N Engl J Med 2009;360:2277-88.

10. Small AJ, Araujo JL, Leggett CL, et al. Radiofrequency ablation is asso-ciated with decreased neoplastic progression in patients withbarrett’s esophagus and confirmed low-grade dysplasia. Gastroenter-ology 2015;149:567-76; quiz e513-64.

11. Qumseya BJ, Wani S, Gendy S, et al. Disease progression in Barrett’slow-grade dysplasia with radiofrequency ablation compared with sur-veillance: systematic review and meta-analysis. Am J Gastroenterol2017;112:849-65.

12. American Gastroenterological Association; Spechler SJ, Sharma P,Souza RF, et al. American Gastroenterological Association medical po-sition statement on the management of Barrett’s esophagus. Gastro-enterology 2011;140:1084-91.

13. Fitzgerald RC, di Pietro M, Ragunath K, et al. British Society of Gastro-enterology guidelines on the diagnosis and management of Barrett’soesophagus. Gut 2014;63:7-42.

14. Weusten B, Bisschops R, Coron E, et al. Endoscopic management ofBarrett’s esophagus: European Society of Gastrointestinal Endoscopy(ESGE) position statement. Endoscopy 2017;49:191-8.

15. Saxena N, Inadomi JM. Effectiveness and cost-effectiveness of endo-scopic screening and surveillance. Gastrointest Endosc Clin North Am2017;27:397-421.

16. Abela JE, Going JJ, Mackenzie JF, et al. Systematic four-quadrant bi-opsy detects Barrett’s dysplasia in more patients than nonsystematicbiopsy. Am J Gastroenterol 2008;103:850-5.

17. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensuson rating quality of evidence and strength of recommendations. BMJ2008;336:924-6.

18. Codipilly DC, Chandar AK, Singh S, et al. The effect of endoscopic sur-veillance in patients with Barrett’s esophagus: a systematic reviewand meta-analysis. Gastroenterology 2018;154:2068-86.

19. El-Serag HB, Naik AD, Duan Z, et al. Surveillance endoscopy is associ-ated with improved outcomes of oesophageal adenocarcinoma de-tected in patients with Barrett’s oesophagus. Gut 2016;65:1252-60.

20. Tramontano AC, Sheehan DF, Yeh JM, et al. The impact of a priordiagnosis of Barrett’s esophagus on esophageal adenocarcinoma sur-vival. Am J Gastroenterol 2017;112:1256-64.

21. ASGE Standards of Practice Committee; Ben-Menachem T, Decker GA,Early DS, et al. Adverse events of upper GI endoscopy. GastrointestEndosc 2012;76:707-18.

22. Qumseya BJ, Wani S, Desai M, et al. Adverse events after radiofre-quency ablation in patients with Barrett’s esophagus: a systematic re-view and meta-analysis. Clin Gastroenterol Hepatol 2016;14:1086-95.

23. Inadomi JM, Sampliner R, Lagergren J, et al. Screening and surveil-lance for Barrett esophagus in high-risk groups: a cost-utility analysis.Ann Intern Med 2003;138:176-86.

24. Nietert PJ, Silverstein MD, Mokhashi MS, et al. Cost-effectiveness ofscreening a population with chronic gastroesophageal reflux. Gastro-intest Endosc 2003;57:311-8.

25. Gerson L, Lin OS. Cost-benefit analysis of capsule endoscopycompared with standard upper endoscopy for the detection of Bar-rett’s esophagus. Clin Gastroenterol Hepatol 2007;5:319-25.

26. Gerson LB, Groeneveld PW, Triadafilopoulos G. Cost-effectivenessmodelof endoscopic screening and surveillance in patients with gastroesoph-ageal reflux disease. Clin Gastroenterol Hepatol 2004;2:868-79.

27. Rubenstein JH, Inadomi JM, Brill JV, et al. Cost utility of screening forBarrett’s esophagus with esophageal capsule endoscopy versus con-ventional upper endoscopy. Clin Gastroenterol Hepatol 2007;5:312-8.

www.giejournal.org

Page 23: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

28. Rubenstein JH, Inadomi JM. Defining a clinically significant adverseimpact of diagnosing Barrett’s esophagus. J Clin Gastroenterol2006;40:109-15.

29. Old O, Moayyedi P, Love S, et al. Barrett’s Oesophagus Surveillanceversus endoscopy at need Study (BOSS): protocol and analysis planfor a multicentre randomized controlled trial. J Med Screen2015;22:158-64.

30. Wani S, Falk G, Hall M, et al. Patients with nondysplastic Barrett’sesophagus have low risks for developing dysplasia or esophagealadenocarcinoma. Clin Gastroenterol Hepatol 2011;9:220-7; quiz e226.

31. Hvid-Jensen F, Pedersen L, Drewes AM, et al. Incidence of adenocar-cinoma among patients with Barrett’s esophagus. N Engl J Med2011;365:1375-83.

32. Desai TK, Krishnan K, Samala N, et al. The incidence of oesophagealadenocarcinoma in non-dysplastic Barrett’s oesophagus: a meta-analysis. Gut 2012;61:970-6.

33. ASGE Standards of Practice Committee; Evans JA, Early DS, Fukami N,et al. The role of endoscopy in Barrett’s esophagus and other prema-lignant conditions of the esophagus. Gastrointest Endosc 2012;76:1087-94.

34. Wani S, Gaddam S. Best practices in surveillance of Barrett’s esoph-agus [Editorial]. Am J Gastroenterol 2017;112:1056-60.

35. Reid BJ, Blount PL, Feng Z, et al. Optimizing endoscopic biopsy detec-tion of early cancers in Barrett’s high-grade dysplasia. Am J Gastroen-terol 2000;95:3089-96.

36. Fitzgerald RC, Saeed IT, Khoo D, et al. Rigorous surveillance protocolincreases detection of curable cancers associated with Barrett’sesophagus. Dig Dis Sci 2001;46:1892-8.

37. Abrams JA, Kapel RC, Lindberg GM, et al. Adherence to biopsy guide-lines for Barrett’s esophagus surveillance in the community setting inthe United States. Clin Gastroenterol Hepatol 2009;7:736-42; quiz 710.

38. Wani S, Keswani RN, Petersen B, et al. Training in EUS and ERCP: stan-dardizing methods to assess competence. Gastrointest Endosc2018;87:1371-82.

39. Wani S, Mathur SC, Curvers WL, et al. Greater interobserver agree-ment by endoscopic mucosal resection than biopsy samples in Bar-rett’s dysplasia. Clin Gastroenterol Hepatol 2010;8:783-8.

40. Wani S, Falk GW, Post J, et al. Risk factors for progression of low-gradedysplasia in patients with Barrett’s esophagus. Gastroenterology2011;141:1179-86.

41. Curvers WL, ten Kate FJ, Krishnadath KK, et al. Low-grade dysplasia inBarrett’s esophagus: overdiagnosed and underestimated. Am J Gas-troenterol 2010;105:1523-30.

42. Krishnamoorthi R, Singh S, Ragunathan K, et al. Factors associatedwith progression of Barrett’s esophagus: a systematic review andmeta-analysis. Clin Gastroenterol Hepatol 2018;16:1046-55.

43. Jankowski JAZ, de Caestecker J, Love SB, et al. Esomeprazole andaspirin in Barrett’s oesophagus (AspECT): a randomised factorial trial.Lancet 2018;392:400-8.

44. Thrift AP, Kendall BJ, Pandeya N, et al. A model to determine absoluterisk for esophageal adenocarcinoma. Clin Gastroenterol Hepatol2013;11:138-44.

45. Bird-Lieberman EL, Dunn JM, Coleman HG, et al. Population-basedstudy reveals new risk-stratification biomarker panel for Barrett’sesophagus. Gastroenterology 2012;143:927-35.

46. Sato F, Jin Z, Schulmann K, et al. Three-tiered risk stratification modelto predict progression in Barrett’s esophagus using epigenetic andclinical features. PLoS One 2008;3:e1890.

47. Parasa S, Vennalaganti S, Gaddam S, et al. Development and valida-tion of a model to determine risk of progression of Barrett’s esoph-agus to neoplasia. Gastroenterology 2018;154:1282-9.

48. Taylor JB, Rubenstein JH. Meta-analyses of the effect of symptoms ofgastroesophageal reflux on the risk of Barrett’s esophagus. Am J Gas-troenterol 2010;105:1729; quiz 1738.

49. Andrici J, Cox MR, Eslick GD. Cigarette smoking and the risk of Barrett’sesophagus: a systematic review and meta-analysis. J GastroenterolHepatol 2013;28:1258-73.

www.giejournal.org

50. Singh S, Sharma AN, Murad MH, et al. Central adiposity is associatedwith increased risk of esophageal inflammation, metaplasia, andadenocarcinoma: a systematic review and meta-analysis. Clin Gastro-enterol Hepatol 2013;11:1399-412.

51. Qumseya BJ, Bukannen A, Gendy S, et al. Systematic review andmeta-analysis of prevalence and risk factors for Barrett’s esophagus.Gastrointest Endosc. Epub 2019 May 29.

52. Gupta N, Bansal A, Wani SB, et al. Endoscopy for upper GI cancerscreening in the general population: a cost-utility analysis. Gastroint-est Endosc 2011;74:610-24.

53. Benaglia T, Sharples LD, Fitzgerald RC, et al. Health benefits and costeffectiveness of endoscopic and nonendoscopic cytospongescreening for Barrett’s esophagus. Gastroenterology 2013;144:62-73.

54. Heberle CR, Omidvari AH, Ali A, et al. Cost effectiveness of screeningpatients with gastroesophageal reflux disease for Barrett’s esophaguswith a minimally invasive cell sampling device. Clin GastroenterolHepatol 2017;15:1397-404.

55. Inadomi JM, Saxena N. Screening and surveillance for Barrett’s esoph-agus: Is it cost-effective? Dig Dis Sci 2018;63:2094-104.

56. Spechler SJ, Katzka DA, Fitzgerald RC. New screening techniques inBarrett’s esophagus: great ideas or great practice? Gastroenterology2018;154:1594-601.

57. Bhat SK, McManus DT, Coleman HG, et al. Oesophageal adenocarci-noma and prior diagnosis of Barrett’s oesophagus: a population-based study. Gut 2015;64:20-5.

58. Chak A, Faulx A, Eng C, et al. Gastroesophageal reflux symptoms inpatients with adenocarcinoma of the esophagus or cardia. Cancer2006;107:2160-6.

59. Rex DK, Cummings OW, Shaw M, et al. Screening for Barrett’s esoph-agus in colonoscopy patients with and without heartburn. Gastroen-terology 2003;125:1670-7.

60. Gerson LB, Shetler K, Triadafilopoulos G. Prevalence of Barrett’sesophagus in asymptomatic individuals. Gastroenterology 2002;123:461-7.

61. Rubenstein JH, Morgenstern H, Appelman H, et al. Prediction of Bar-rett’s esophagus among men. Am J Gastroenterol 2013;108:353-62.

62. Ireland CJ, Gordon AL, Thompson SK, et al. Validation of a risk predic-tion model for Barrett’s esophagus in an Australian population. ClinExp Gastroenterol 2018;11:135-42.

63. Ireland CJ, Fielder AL, Thompson SK, et al. Development of a risk pre-diction model for Barrett’s esophagus in an Australian population. DisEsophagus 2017;30:1-8.

64. Thrift AP, Kendall BJ, Pandeya N, et al. A clinical risk prediction modelfor Barrett esophagus. Cancer Prev Res (Phila) 2012;5:1115-23.

65. Rubenstein JH, Thrift AP. Risk factors and populations at risk: selectionof patients for screening for Barrett’s oesophagus. Best Pract Res ClinGastroenterol 2015;29:41-50.

66. Mussetto A, Manno M, Fuccio L, et al. Screening for Barrett’s oesoph-agus with oesophageal capsule endoscopy in first-degree relatives ofpatients affected by Barrett’s oesophagus: results of a pilot study.Arab J Gastroenterol 2013;14:51-4.

67. Galmiche JP, Sacher-Huvelin S, Coron E, et al. Screening for esopha-gitis and Barrett’s esophagus with wireless esophageal capsuleendoscopy: a multicenter prospective trial in patients with refluxsymptoms. Am J Gastroenterol 2008;103:538-45.

68. Chavalitdhamrong D, Chen GC, Roth BE, et al. Esophageal capsule endos-copy for evaluation of patients with chronic gastroesophageal refluxsymptoms: findings and its image quality. Dis Esophagus 2011;24:295-8.

69. Atkinson M, Das A, Faulx A, et al. Ultrathin esophagoscopy inscreening for Barrett’s esophagus at a Veterans Administration Hospi-tal: easy access does not lead to referrals. Am J Gastroenterol2008;103:92-7.

70. Juhasz A, Mittal SK, Lee TH, et al. Prevalence of Barrett esophagus infirst-degree relatives of patients with esophageal adenocarcinoma.J Clin Gastroenterol 2011;45:867-71.

71. Qumseya BJ, Wang H, Badie N, et al. Advanced imaging technologiesincrease detection of dysplasia and neoplasia in patients with

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 357

Page 24: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

Barrett’s esophagus: a meta-analysis and systematic review. Clin Gas-troenterol Hepatol 2013;11:1562-70.

72. ASGE Technology Committee; Thosani N, Abu Dayyeh BK, Sharma P,et al. ASGE Technology Committee systematic review and meta-analysis assessing the ASGE Preservation and Incorporation of ValuableEndoscopic Innovations thresholds for adopting real-time imaging-as-sisted endoscopic targeted biopsy during endoscopic surveillance ofBarrett’s esophagus. Gastrointest Endosc 2016;83:684-98.

73. Sharma P, Savides TJ, Canto MI, et al. The American Society forGastrointestinal Endoscopy PIVI (Preservation and Incorporation ofValuable Endoscopic Innovations) on imaging in Barrett’s esophagus.Gastrointest Endosc 2012;76:252-4.

74. Bratlie SO, Johnsson E, Jonsson C, et al. Multiple-band imaging pro-vides better value than white-light endoscopy in detection ofdysplasia in patients with Barrett’s esophagus. Clin GastroenterolHepatol 2015;13:1068-74.

75. Curvers WL, van Vilsteren FG, Baak LC, et al. Endoscopic trimodal im-aging versus standard video endoscopy for detection of early Bar-rett’s neoplasia: a multicenter, randomized, crossover study ingeneral practice. Gastrointest Endosc 2011;73:195-203.

76. Sharma P, Hawes RH, Bansal A, et al. Standard endoscopy withrandom biopsies versus narrow band imaging targeted biopsies inBarrett’s oesophagus: a prospective, international, randomisedcontrolled trial. Gut 2013;62:15-21.

77. Bhandari P, Kandaswamy P, Cowlishaw D, et al. Acetic acid-enhancedchromoendoscopy is more cost-effective than protocol-guided biopsiesin a high-risk Barrett’s population. Dis Esophagus 2012;25:386-92.

78. Olliver JR, Wild CP, Sahay P, et al. Chromoendoscopy with methyleneblue and associated DNA damage in Barrett’s oesophagus. Lancet2003;362:373-4.

79. Kondo H, Fukuda H, Ono H, et al. Sodium thiosulfate solution sprayfor relief of irritation caused by Lugol’s stain in chromoendoscopy.Gastrointest Endosc 2001;53:199-202.

80. Wani S, Sharma P. Endoscopic surface imaging of Barrett’s esophagus:an optimistic view. Gastroenterology 2007;133:11-3.

81. Sharma P, Bergman JJ, Goda K, et al. Development and validation of aclassification system to identify high-grade dysplasia and esophagealadenocarcinoma in Barrett’s esophagus using narrow-band imaging.Gastroenterology 2016;150:591-8.

82. Everson MA, Lovat LB, Graham DG, et al. Virtual chromoendoscopyusing optical enhancement improves the detection of Barrett’sesophagus-associated neoplasia. Gastrointest Endosc 2019;89:247-56.

83. Xiong YQ, Ma SJ, Zhou JH, et al. A meta-analysis of confocal laser en-domicroscopy for the detection of neoplasia in patients with Barrett’sesophagus. J Gastroenterol Hepatol 2016;31:1102-10.

84. Canto MI, Anandasabapathy S, Brugge W, et al. In vivo endomicro-scopy improves detection of Barrett’s esophagus-related neoplasia:a multicenter international randomized controlled trial (with video).Gastrointest Endosc 2014;79:211-21.

85. Dunbar KB, Canto MI. Confocal laser endomicroscopy in Barrett’sesophagus and endoscopically inapparent Barrett’s neoplasia: a pro-spective, randomized, double-blind, controlled, crossover trial. Gas-trointest Endosc 2010;72:668.

86. Sharma P, Meining AR, Coron E, et al. Real-time increased detection ofneoplastic tissue in Barrett’s esophagus with probe-based confocallaser endomicroscopy: final results of an international multicenter,prospective, randomized, controlled trial. Gastrointest Endosc2011;74:465-72.

87. Bertani H, Frazzoni M, Dabizzi E, et al. Improved detection of incidentdysplasia by probe-based confocal laser endomicroscopy in a Bar-rett’s esophagus surveillance program. Dig Dis Sci 2013;58:188-93.

88. Wallace MB, Meining A, Canto MI, et al. The safety of intravenous fluo-rescein for confocal laser endomicroscopy in the gastrointestinaltract. Aliment Pharmacol Ther 2010;31:548-52.

89. ASGE Technology Committee; Chauhan SS, Abu Dayyeh BK, Bhat YM,et al. Confocal laser endomicroscopy. Gastrointest Endosc 2014;80:928-38.

358 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3 : 2019

90. Saumoy M, Schneider Y, Novikov AA, et al. Cost-utility analysis for sur-veillance of Barrett’s esophagus evaluating standard biopsiescompared to optical coherence tomography and confocal laser endo-microscopy [abstract]. Gastrointest Endosc 2017;85:AB559-60.

91. Machicado JD, Han S, Yadlapati RH, et al. A survey of expert practiceand attitudes regarding advanced imaging modalities in surveillanceof Barrett’s esophagus. Dig Dis Sci 2018;63:3262-71.

92. Qumseya BJ, Brown J, Abraham M, et al. Diagnostic performance ofEUS in predicting advanced cancer among patients with Barrett’sesophagus and high-grade dysplasia/early adenocarcinoma: system-atic review and meta-analysis. Gastrointest Endosc 2015;81:865-74.

93. Qumseya BJ, Bartel MJ, Gendy S, et al. High rate of over-staging ofBarrett’s neoplasia with endoscopic ultrasound: Systemic reviewand meta-analysis. Dig Liver Dis 2018;50:438-45.

94. Thota PN, Sada A, Sanaka MR, et al. Correlation between endoscopicforceps biopsies and endoscopic mucosal resection with endoscopicultrasound in patients with Barrett’s esophagus with high-gradedysplasia and early cancer. Surg Endosc 2017;31:1336-41.

95. Bartel MJ, Wallace TM, Gomez-Esquivel RD, et al. Role of EUS in pa-tients with suspected Barrett’s esophagus with high-grade dysplasiaor early esophageal adenocarcinoma: impact on endoscopic therapy.Gastrointest Endosc 2017;86:292-8.

96. Yousef F, Cardwell C, Cantwell MM, et al. The incidence of esophagealcancer and high-grade dysplasia in Barrett’s esophagus: a systematicreview and meta-analysis. Am J Epidemiol 2008;168:237-49.

97. Eloubeidi MA, Tamhane A, Lopes TL, et al. Cervical esophageal perfo-rations at the time of endoscopic ultrasound: a prospective evalua-tion of frequency, outcomes, and patient management. Am JGastroenterol 2009;104:53-6.

98. Das A, Sivak MV Jr, Chak A. Cervical esophageal perforation duringEUS: a national survey. Gastrointest Endosc 2001;53:599-602.

99. Wang KX, Ben QW, Jin ZD, et al. Assessment of morbidity and mortal-ity associated with EUS-guided FNA: a systematic review. GastrointestEndosc 2011;73:283-90.

100. Yokoyama K, Ushio J, Numao N, et al. Esophageal seeding after endo-scopic ultrasound-guided fine-needle aspiration of a mediastinal tu-mor. Endosc Int Open 2017;5:E913-7.

101. Alzoubaidi D, Ragunath K, Wani S, et al. Quality indicators for Barrett’sendotherapy (QBET): UK consensus statements for patients undergo-ing endoscopic therapy for Barrett’s neoplasia. United Eur Gastroent2018;6(Suppl 1):A37.

102. Anandasabapathy S, Sontag S, Graham DY, et al. Computer-assistedbrush-biopsy analysis for the detection of dysplasia in a high-riskBarrett’s esophagus surveillance population. Dig Dis Sci 2011;56:761-6.

103. Vennalaganti PR, Kaul V, Wang KK, et al. Increased detection ofBarrett’s esophagus-associated neoplasia using wide-area trans-epithelial sampling: a multicenter, prospective, randomized trial. Gas-trointest Endosc 2018;87:348-55.

104. Iorio N, Sprung B, Kaul V, et al. Transepithelial brush biopsy withcomputer-assisted tissue analysis increases detection of residual orrecurrent intestinal metaplasia and dysplasia following endoscopicablation of Barrett’s esophagus [abstract]. Gastrointest Endosc2015;81:AB141.

105. Smith MS, Ikonomi E, Bhuta R, et al. Wide-area transepithelial sam-pling with computer-assisted 3-dimensional analysis (WATS) mark-edly improves detection of esophageal dysplasia and Barrett’sesophagus: analysis from a prospective multicenter community-based study. Dis Esophagus 2019;32:doy099.

106. Gross SA, Smith MS, Kaul V, et al. Increased detection of Barrett’sesophagus and esophageal dysplasia with adjunctive use of wide-area transepithelial sample with three-dimensional computer-assisted analysis (WATS). United Eur Gastroenterol J 2018;6:529-35.

107. Johanson JF, Frakes J, Eisen D, Endo CCG. Computer-assisted analysisof abrasive transepithelial brush biopsies increases the effectivenessof esophageal screening: a multicenter prospective clinical trial bythe EndoCDx Collaborative Group. Dig Dis Sci 2011;56:767-72.

www.giejournal.org

Page 25: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

108. Smith M, Iorio N, Walzer E, et al. Wide area transepithelial samplingwith computer-assisted 3-dimensional analysis (WATS(3D)) safelyevaluates a variety of esophageal disorders. Am J Gastroenterol2014:S24.

109. Shaheen NJ, Smith M, Goldblum J, et al. Progression of Barrett’s esoph-agus (BE) and dysplasia detected by wide area transepithelial samplingwith computer-assisted 3D analysis (WATS3D) confirms the clinical sig-nificance of crypt dysplasia. Am J Gastroenterol 2018;113:s172.

110. Qumseya BJ, Gendy S, Qumsiyeh Y, et al. marginal increase indysplasia detection and very high false positive rate for volumetriclaser endomicroscopy in Barrett’s esophagus: systemic review andmeta-analysis [abstract]. Gastrointest Endosc 2017;85:AB554.

111. Alshelleh M, Inamdar S, McKinley M, et al. Incremental yield ofdysplasia detection in Barrett’s esophagus using volumetric laser en-domicroscopy with and without laser marking compared with a stan-dardized random biopsy protocol. Gastrointest Endosc 2018;88:35-42.

112. Wolfsen HC, Sharma P, Wallace MB, et al. Safety and feasibility ofvolumetric laser endomicroscopy in patients with Barrett’s esophagus(with videos). Gastrointest Endosc 2015;82:631-40.

113. van der Sommen F, Curvers WL, Nagengast WB. Novel develop-ments in endoscopic mucosal imaging. Gastroenterology 2018;154:1876-86.

114. Swager AF, Tearney GJ, Leggett CL, et al. Identification of volumetriclaser endomicroscopy features predictive for early neoplasia in Bar-rett’s esophagus using high-quality histological correlation. Gastroint-est Endosc 2017;85:918-26.

115. Leggett CL, Gorospe EC, Chan DK, et al. Comparative diagnostic per-formance of volumetric laser endomicroscopy and confocal laser en-domicroscopy in the detection of dysplasia associated with Barrett’sesophagus. Gastrointest Endosc 2016;83:880-8.

116. Swager AF, van der Sommen F, Klomp SR, et al. Computer-aideddetection of early Barrett’s neoplasia using volumetric laser endomi-croscopy. Gastrointest Endosc 2017;86:839-46.

117. Dong J, Buas MF, Gharahkhani P, et al. Determining risk ofBarrett’s esophagus and esophageal adenocarcinoma based onepidemiologic factors and genetic variants. Gastroenterology2018;154:1273-81.

118. Thota PN, Chak A. Mass screening for Barrett’s esophagus: myth orreality? Clin Gastroenterol Hepatol 2019;17:610-2.

119. Cook MB, Wild CP, Forman D. A systematic review and meta-analysisof the sex ratio for Barrett’s esophagus, erosive reflux disease, andnonerosive reflux disease. Am J Epidemiol 2005;16211:1050-61.

Abbreviations: AE, adverse event; ASGE, American Society forGastrointestinal Endoscopy; BE, Barrett’s esophagus; CE,chromoendoscopy; CI, confidence interval; CLE, confocal laserendomicroscopy; EAC, esophageal adenocarcinoma; EET, endoscopiceradication therapies; GRADE, Grading of RecommendationsAssessment, Development and Evaluation; HGD, high-grade dysplasia;HR, hazard ratio; LGD, low-grade dysplasia; MA, meta-analysis; NBI,narrow-band imaging; NDBE, nondysplastic Barrett’s esophagus; OR,odds ratio; PICO, Population, Intervention, Comparator andOutcomes; PIVI, Preservation and Incorporation of Valuable

www.giejournal.org

Endoscopic Innovations; RCT, randomized controlled trial; RR,relative risk; SR, systematic review; VC, virtual chromoendoscopy; VLE,volumetric laser endomicroscopy; WATS, wide-area transepithelialsampling; WATS-3D, WATS with computer-assisted 3-dimensionalanalysis; WLE, white-light endoscopy.

DISCLOSURE: The following authors disclosed financial relationships relevantto this publication: M. Al-Haddad: Teaching and research support fromBoston Scientific. J. Buxbaum: Consultant for Olympus. J. Dewitt:Consultant for Olympus America and Boston Scientific; grant support fromPinnacle Biologics. D. Fishman: Contributor to UpToDate. B. Jacobson:Consultant for Motus GI, Remedy Partners, and Dark Canyon, LLC. L. Jamil:Consultant and speaker for Aries Pharmaceutical. M. Khashab: Consultantand medical advisory board for Boston Scientific and Olympus; consultantfor Medtronic. M. McCarter: Medical advisory board for Debbies DreamFoundation. N. Thosani: Consultant for Pentax America, Boston Scientific,Medtronic, and Endogastric Solutions; speaker for AbbVie. S. Wani:Consultant for Medtronic and Boston Scientific; supported by University ofColorado Department of Medicine Outstanding Early Scholars Program. Allother authors disclosed no financial relationships relevant to this publication.

*Drs Qumseya and Sultan contributed equally to this article.

Copyright ª 2019 by the American Society for Gastrointestinal Endoscopy0016-5107/$36.00https://doi.org/10.1016/j.gie.2019.05.012

Received May 6, 2019. Accepted May 6, 2019.

Current affiliations: Department of Gastroenterology, Archbold MedicalGroup, Thomasville, Georgia, USA (1), Division of Gastroenterology, Hepa-tology and Nutrition, University of Minnesota, Minneapolis, Minnesota,USA (2), Harvard School of Public Health, Boston, Massachusetts, USA(3), Pancreatic and Biliary Diseases Program, Cedars-Sinai Medical Center,Los Angeles, California, USA (4), Boston University, Boston, Massachusetts,USA (5), Baylor Global Initiatives, Baylor College of Medicine, Houston,Texas, USA (6), Division of Digestive and Liver Diseases, University of TexasSouthwestern Medical Center, Dallas, Texas, USA (7), Division of Gastroin-testinal and Liver Diseases, Keck School of Medicine, University of South-ern California, Los Angeles, California, USA (8), Department ofGastroenterology, Baylor College of Medicine, Texas Children’s Hospital,Houston, Texas, USA, (9) Department of Gastroenterology, Mayo Clinic Ari-zona, Scottsdale, Arizona, USA (10), The Permanente Medical Group, KaiserPermanente San Francisco Medical Center, San Francisco, California, USA(11), Division of General Internal Medicine and Public Health, VanderbiltUniversity Medical Center, Nashville, Tennessee, USA (12), Department ofGastroenterology, Kaiser Permanente San Francisco Medical Center, SanFrancisco, California, USA (13), Division of Gastroenterology and Hepatol-ogy, Johns Hopkins University, Baltimore, Maryland, USA (14), Division ofGastroenterology and Hepatology, University of Iowa Hospitals & Clinics,Iowa City, Iowa, USA (15), Division of Gastroenterology, Hepatology andNutrition, McGovern Medical School, UTHealth, Houston, Texas, USA (16),Division of Gastroenterology, Montefiore Medical Center, Albert EinsteinCollege of Medicine, Bronx, New York, USA (17), Indiana University MedicalCenter, Carmel, Indiana, USA (18), Division of Gastroenterology and Hepa-tology, University of Colorado Anschutz Medical Center, Aurora, Colorado,USA (19).

Volume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 359

Page 26: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

APPENDIX 1. Search strategies for the PICOquestion 2 (screening), 3 (update CE), 4(update CLE), 5

Search strategies screening for Barrett’s esophagus(PICO 2)

Date of last Search: 09/06/2017

Medline

359.e1 GASTROINTESTINAL ENDOSCOPY Volume 90, No. 3

1387

Embase 1610 Web of Science 1235 Total 4232

After removing duplicates 2502 records

MEDLINE (Ovid)Ovid MEDLINE(R) Epub Ahead of Print, In-Process &

Other Non-Indexed Citations, Ovid MEDLINE(R) Daily,Ovid MEDLINE and Versions(R)

201709051387 Records(exp Endoscopy, Digestive System/ OR (esophagogas-

troduodenoscop* OR egd OR oesophagogastroduodeno-scop* OR ogd).ab,ti OR ((oesophago gastro OResophago gastro) adj duodenoscop*).ab,ti OR ((screen*OR surveill* OR upper) adj3 endoscop*).ab,ti) AND (expmass screening/ OR exp early detection of cancer/exp OR(screen* OR surveillance).ab,ti)

AND (exp barrett esophagus/ OR ((esophag* OR oeso-phag*) AND barrett*).ab,ti)

Embase (Elsevier)201709051610 Records(’esophagogastroduodenoscopy’/exp OR (esophagogas-

troduodenoscop* OR egd OR oesophagogastroduodeno-scop* OR ogd):ab,ti OR ((’oesophago-gastro’ OR’esophago gastro’) NEXT/1 duodenoscop*):ab,ti OR((screen* OR surveill* OR upper) NEAR/3 endosco-p*):ab,ti) AND (’mass screening’/exp OR ’early cancer diag-nosis’/exp OR (screen* OR surveillance):ab,ti) AND(’Barrett esophagus’/exp OR ((esophag* OR oesophag*)AND barrett*):ab,ti)

Web of Science (Clarivate Analytics)IndexesZSCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-

SSH, BKCI-S, BKCI-SSH, ESCI, CCR-EXPANDED, IC Time-spanZAll years

201709051235 Records

TSZ("esophagogastroduodenoscop*" OR "egd" OR "oe-sophagogastroduodenoscop*" OR "ogd" OR "oesophagogastro duodenoscop*" OR "esophago gastro duodeno-scop*" OR (("screen*" OR "surveill*" OR "upper") NEAR/3"endoscop*")) AND TSZ("screen*" OR "surveillance")AND TSZ(("esophag*" OR "oesophag*") AND "barrett*")

: 2019

Search strategies forupdateonchromoendoscopyfor Barrett’s esophagus (PICO 3)

Date of original search: 10/1/2012This search was run on 06/12/2017In both PubMed and Embase, the search was limited by

the date the record was entered into the database (the en-trez date, or EDAT) set to 20120101 (Jan. 1, 2012). For Webof Science, the publication year was used and set to 2012.

PubMed

www.giejourna

493

Embase 983 Web of Science 458 Total 1934

1354 Records after removing duplicates.PubMed (NCBI)20170612493 Records((Esophagoscopy[mesh] AND Image Enhancement

[mesh]) OR "Acetic Acid"[Mesh] OR "Indigo Carmine"[Mesh] OR "Methylene Blue"[Mesh] OR "Congo Red"[mesh] OR "Gentian Violet"[mesh] OR "Phenolsulfonph-thalein"[mesh] OR "Coloring Agents"[Mesh] OR "Chromo-genic Compounds"[mesh] OR "Iodides"[mesh] OR aceticacid[tiab] OR indigo[tiab] OR carmine[tiab] OR methyleneblue[tiab] OR phenolsulfonphthalein[tiab] OR congo red[tiab] OR gentian violet[tiab] OR phenol red[tiab] OR crys-tal violet[tiab] OR lugol*[tiab] OR iodine[tiab] OR coloringagent*[tiab] OR colouring agent*[tiab] OR chromoendo-scop*[tiab] OR chromo endoscop*[tiab] OR chromo-scop*[tiab] OR narrow band imag*[tiab] OR nbi[tiab] ORautofluoresc*[tiab] OR trimodal[tiab] OR fujinon intelli-gent[tiab]) AND ("Barrett Esophagus"[Mesh] OR "Esopha-geal Neoplasms"[Mesh] OR ((esophag*[tiab] ORoesophag*) AND (neoplasm*[tiab] OR neoplasia*[tiab]OR cancer*[tiab] OR dysplas*[tiab] OR carcinoma*[tiab]OR precancer*[tiab] OR metaplas*[tiab] OR barrett*[tiab]))) AND ("2012/01/01"[EDAT] : "3000/12/31"[EDAT])

Embase (Elsevier)20170612983 records(’chromoendoscopy’/exp OR ’acetic acid’/exp OR ’in-

digo carmine’/exp OR ’methylene blue’/exp OR ’congored’/exp OR ’crystal violet’/exp OR ’phenolsulfonphtha-lein’/exp OR ’coloring agent’/exp OR ’chromogenic sub-strate’/exp OR ’iodide’/exp OR ’acetic acid’:ab,ti ORindigo:ab,ti OR carmine:ab,ti OR ’methylene blue’:ab,tiOR phenolsulfonphthalein:ab,ti OR ’congo red’:ab,tiOR ’gentian violet’:ab,ti OR ’phenol red’:ab,ti OR ’crystalviolet’:ab,ti OR lugol*:ab,ti OR iodine:ab,ti OR ’coloringagent*’:ab,ti OR ’colouring agent*’:ab,ti OR chromoendo-scop*:ab,ti OR (chromo NEXT/1 endoscop*):ab,ti OR chro-moscop*:ab,ti OR (’narrow band’ NEXT/1 image*):ab,ti ORnbi:ab,ti OR autofluoresc*:ab,ti OR trimodal:ab,ti OR

l.org

Page 27: ASGE guideline on screening and surveillance of Barrett's ...

ASGE guideline on screening and surveillance of BE

’fujinon intelligent’:ab,ti) AND (’Barrett esophagus’/exp OR’esophagus tumor’/exp OR ((esophag* OR oesophag*)AND (neoplasm* OR neoplasia* OR cancer* OR dysplas*OR carcinoma* OR precancer* OR metaplas* OR barret-t*)):ab,ti) AND

[1-1-2012]/sd NOT [31-12-2017]/sd

Web of Science (Thomson Reuters)20170612458 Records

TSZ("acetic acid" OR "indigo" OR "carmine" OR "methy-lene blue" OR "phenolsulfonphthalein" OR "congo red" OR"gentian violet" OR "phenol red" OR "crystal violet" OR"lugol*" OR "iodine" OR "coloring agent*" OR "colouringagent*" OR "chromoendoscop*" OR "chromo endoscop*"OR "chromoscop*" OR "narrow band image*" OR "nbi" OR"autofluoresc*" OR "trimodal" OR "fujinon intelligent")ANDTSZ(("esophag*"OR "oesophag*") AND ("neoplasm*"OR "neoplasia*" OR "cancer*" OR "dysplas*" OR "carci-noma*" OR "precancer*" OR "metaplas*" OR "barrett*"))

Search strategy for update on confocal laserendomicroscopy in Barrett’s esophagus

Date of last search: 06/20/2017Previous search578 Records on 09/06/2016This Update

PubMed

www.giejournal.org

226

Embase 479 Web of Science 396 BIOSIS Total 1101 After removing duplicates 778

After substracting original search, 205 new records re-mained; 5 old records did not match records in the newsearch

Vo

PubMed (NCBI)20170620226 Records (truncated "Barrett", added "dysplasia*";

157 Records previously)

("Microscopy, Confocal"[mesh] OR (confocal[tiab] AND(endomicroscop*[tiab] OR mircoscop*[tiab])) OR cle[tiab] OR pcle[tiab]) AND ("Barrett Esophagus"[mesh]OR (barrett*[tiab] AND (esophag*[tiab] OR oesophag*[tiab] OR neoplas*[tiab] OR dysplasia*[tiab])))

Embase (Elsevier)20170620479 Records

(’confocal laser microscopy’/exp OR (confocal NEAR/3(endomicroscop* OR microscop*)):ab,ti OR cle:ab,ti ORpcle:ab,ti) AND (’barrett esophagus’/exp OR (barrett*AND (esophag* OR oesophag* OR neoplas* ORdysplasia*)):ab,ti)

Web of Science (Thomson Reuters)IndexesZSCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-

SSH, BKCI-S, BKCI-SSH, ESCI, CCR-EXPANDED, IC Time-spanZAll years

20170620396 Records

TSZ(("confocal" NEAR/3 ("endomicroscop*" OR "micro-scop*")) OR "cle" OR "pcle") AND TSZ("barrett*" AND("esophag*" OR "oesophag*" OR "neoplas*" OR"dysplasia*"))

BIOSIS Previews (Thomson Reuters)20170620176 Records

TSZ(("confocal" NEAR/3 ("endomicroscop*" OR "micro-scop*")) OR "cle" OR "pcle") AND TSZ("barrett*" AND("esophag*" OR "oesophag*" OR "neoplas*" OR"dysplasia*"))

lume 90, No. 3 : 2019 GASTROINTESTINAL ENDOSCOPY 359.e2